23: Twitter’s Bias Bounty Program and The Ethics of Research with Dr. Borhane Blili-Hamelin

Seth Villegas
Okay say something again.

Ryan Mayfield
Hello, Hello?

Seth Villegas
There we go. Okay.

Ryan Mayfield
Okay!

Seth Villegas
I moved this thing like one millimeter and it didn’t work anyway. Another great example, if you depend on technology for everything- If a wire needs to be at a specific angle for it to work, it doesn’t inspire a lot of confidence in me personally.

Ryan Mayfield
No, it does not.

Seth Villegas
Yeah, but thank you, Ryan, so much for being here. I always like to admit that there are technical difficulties on a podcast about technology, because there’s always technical difficulties. So if you assume that everything works as intended, well, maybe it won’t.

Ryan Mayfield
It wouldn’t be very resilient of you to assume that technology will always work.

Seth Villegas
Well, you wouldn’t need technology to be resilient if it did work, which again, always brings me back to my favorite line in the movie Spaceballs, which is that even in the future, nothing works. So we met each other at Stanford, a little bit about your background in political science, someone who’s very interested in safety and defense. And so it’d be really great to hear just a little bit about your background, the kinds of things that you’ve been up to, and how that’s got you to the point where you’re at now.

Ryan Mayfield
Yeah, well, thanks for having me, thrilled to be here. So my passion is really physical safety and security issues. And this goes back to my earliest days, the first book I ever remember reading when I was five was about the Battle of Midway. So I’ve been doing this for a long time. My mom was a VA nurse. And so I grew up just begging her to take me to the facility where she works, so I can just shoot the breeze with veterans. And this was, you know, pre Iraq, pre Afghanistan. So you had a lot of mostly Vietnam vets. And so obviously, that brought in a lot of feelings about counterinsurgency, and really got me focused on the human side of conflict. So much of defense tends to revolve around technology and tanks and aircraft carriers and whatnot. But at the end of the day, it’s really about people. It’s about fear. You know, you talk about the- the drivers of conflict, we like to ascribe that Oh, the other side is evil, or whatever it is, but that really masks over their humanity and masks over that in any conflict, both sides have objectives and aims that they’re trying to meet beyond just messing things up. And so I got to Stanford kind of came from small town, Washington State. So it was a bit of a culture shock there. And I was lucky enough to find myself in the center for national security and cooperation there, which I always say was a bit like my mom’s facility, just people are smarter, you know. And so you had a lot of military officers and diplomats and whatnot. And they were trying to dig into these and I was lucky enough to find a retired Army Colonel, Special Forces Colonel there was doing a lot of work on counterinsurgency and counterterrorism. You know, this was the days of the Iraq drawdown and the Afghan surge and whatnot. So really trying to understand how do you do this right. And the best place on earth to study counterinsurgency and counterterrorism with the Philippines. And so we really focused in on what was going on Philippines, there’s multiple different rebel groups there, lots of great data. I actually got the chance to do some field research in the Philippines while I was in college, and then actually move there after college, continue some of that research, and I even got a chance to interview rebels who had left the different movements, really understand life to their eyes. And those conversations were eye opening to me and took this human side of conflict. And further, it forced me to realize that the other side has a voice and the world isn’t black and white, but many shades of grey. You know, we were kind of raised in the post 9/11 generation of like, you’re either for us or against us. But no one ever asked what that actually meant with it. And you know, talking to these former rebels, I had to admit that if I had lived their lives, I could have very easily have ended up in their shoes, or much worse. The things that they had gone through, the roads they had walked, I was actually impressed by how well they ended up. I was like, I would have walked your shoes, I would have been, you know, dead on the side of the road a long time ago. And the other big thing of, you know, getting the chance to understand their paths and whatnot, is really recognizing that a lot of the drivers of instability are low severity events. So you know, we spend so much money on big security events, on counterterrorism and TSA and bomb threats and active shooters and all those things. And I don’t want to say those aren’t important. Obviously, those are very important. But honestly, if I had the choice, if I had a magic wand, and I could read the world of all slippery sidewalks or all terrorist attacks, it’s an absolute no brainer in favor of slippery sidewalks. You could have so much of a bigger impact on humanity and society by eliminating those low severity events that then have trickle down effects of injuries and chronic pain and economic instability and all those sorts of things. And there’s actually this great project in Minnesota, I believe it’s University of St. Thomas, I think, called the Violence Project. And they’ve actually built the best dataset of active shooter events, and really show heavily that A) vast majority of active shooters actually try to seek help before their events and big tie into mental health concerns and wellness concerns with it as well. And so if you can better have a resilient society that will pick up on these people showing warning signs, seeking help, and actually help them, that’s really how you stop active shooters. And so, you know, I hear about Wi Fi routers that detect gunshots and whatnot. It’s like hey, you know, once once bullets start flying, like things are pretty bad. And so you know, what are we really trying to do to focus in on these drivers of instability, these pieces that create a lot of fear that create a ripple effects for society? And so I really think that we need to spend a lot more emphasis on these high frequency, low severity events and less effort on these low frequency, high severity events.

Seth Villegas
Sure, and I think definitely just start to get more of a picture of what you’re talking about. So if I’m taking anything in of like a high severity event, lots of things I can think of would be something like maybe a really devastating hurricane or a tsunami or something like that that comes through and completely wipes out an area. I think, you know, when you mentioned a place like the Philippines, that definitely comes to mind, but specifically in the Philippines, what do you mean by a kind of low severity events? Are you talking about the kinds of I don’t know, social and economic conditions? Like what what is it that’s causing the stress there, exactly?

Ryan Mayfield
Yeah, yes. So one of the big things that got me keyed into these low severity events is actually working on a tip line that was being used in counterinsurgency campaigns. And we noticed a couple of villages that were kind of breaking the tip line. And were actually using it to deal with more what I would consider these, you know, baseline resilience sorts of things. So instead of it being so focused on like, oh, there might be a rebel, right? coming through, or like there might be a government right coming through, because, you know, bullet to bullet, it was more about, hey, the creek is running a little bit high, or the bridge feels a little unstable, or, Hey, that little lady at the edge of town, I haven’t seen her a couple days, can someone go check on her. And the thing is, is that if the bridge is a little unstable, and the typhoon comes through, you get cut off. Or if the creek is running a little high, and you’re not looking out for the people who live next to the creek, they get displaced. And so it’s more of those sort of baseline resilience things, and then can magnify the effect of these big natural disasters, or can be exploited by groups that are seeking grievances. And we actually saw some interesting things. I was in the Philippines during typhoon Haiyan. So at the time, 2013, it was the strongest storm to ever make landfall and just absolutely destroyed a bunch of the Central Visayas region in the Philippines. And there were actually movements by rebel groups afterwards, to seize on the instability of this, because you already had really fragile infrastructure, really fragile telecommunications, huge amounts of displaced people, some of whom had still been displaced from the previous storm, that that really is ripe for that. And so it’s one of those things where treating these baseline resiliency pieces have a huge impact on minimizing that risk. And then when bigger events happen, I think that’s where you really get to resilience. And so there’s this debate in a lot of the safety and security community about prevention, like, can you prevent events? And I think a lot of times individuals really focus on the end of the timeline, you know, so can we prevent this active shooter? And my philosophy is yes, if we focus less on the active shooter, and we will look at the broader timeline that led up to that event, there were lots of opportunities, we could have gotten involved. And it’s not necessary that we prevented a, you know, someone from experiencing a mental strife or that we prevented harassment or anything like that, but we responded proactively to those events. And we were able to resolve those. And by resolving those more minor events sooner and more efficiently, that’s when you can really kind of cut the timeline and redirect the timeline to then prevent those more extreme events.

Seth Villegas
So to kind of highlight your perspective then, it sounds like what you’re really talking about is it is possible to kind of draw some sort of a narrative line between these, kind of just saying, these low level things that are going on, say, in a village, you know, concerns you have make you a little bit more anxious, the sorts of things that if they accumulate over time will cause someone to be much more likely to be sympathetic to insurgents, maybe become an insurgent themselves.

Ryan Mayfield
Yeah. And then this is where I really then hold that out of a- I feel like it’s easy to kind of typecast like, oh, an impoverished village in some developing country with that, but those exact same things are happening here. Those exact same things are happening in developed countries, where you see a slip and fall, then someone doesn’t have insurance, and then it has cascading effects, pulling people out of school, placing additional stressors, you know, etc, etc. And so one of the big things is, I first moved to the Philippines, working hand in hand with the defense community. And I continued for several years thereafter, I’m working on defense technology. But the thing is, is I mean, the defense community is less focused on resilience, right? I mean, they are focused on big capital, letter S, Security, international security, with that, preventing world war three, all those things. And so I really started to switch over in the technology community and the growing field of safety technology, and because I recognize that, you know, we’ll get this microcosm of these couple of villages that we’ve been studying, but I realized that the same or maybe not the same, exactly, but the similar vibe that was coming off of that was causing a lot of strife back home. And we’ve seen a lot of political strife, a lot of social strife with that. And a lot of these can be traced back down to these baseline resiliency issues, that then are having bigger reverberations across the economy, across society with that, and we can talk a little bit more about about fear versus resilience later if you’d like. And big task of the safety community right now is recognizing that even though by almost any metric, we have never been safer, humans have never been better off, there is this profound amount of anxiety and fear, which is actually then causing us or correlation-causation, all that good stuff, is tied to to greater instability and that you actually have fear taking stable areas and making them more correlated with them becoming less stable through almost nothing other than the influx of fear. And I think there are some elements of the safety community that are maybe trying to make it better, but are indirectly stoking this with it. And then there are other groups that are kind of like hold on Timeout, timeout, timeout, if our efforts are actually making things less stable than like, what are we doing here?

Seth Villegas
A situation I just thought about is, say, the application of drones to taking out an active shooter or using a robot to talk to, you know, again, like a bomb situation, or something. The reason I’m bringing that up, though, is because, I think that your hope seems to be preventing those situations from happening altogether, rather than having some sort of flashy, like, super technological end to that situation that, like something you’d see out of a science fiction movie, but but rather trying to make sure that there are these sorts of different intervening points, or at least creating the kinds of intervening points that you might have to get it so that you’re not in that kind of unsafe situation to begin with? Do you think that’s kind of a fair characterization of what I’m hearing from you?

Ryan Mayfield
So I think it’s really taking a risk based approach to this. Risk is this combination of frequency and severity. So you have these extreme events that is, you know, a bomb going off somewhere. Yes, devastating event with that, but extraordinarily unlikely. And then you have these other events where, something like two times as many people are killed by accidents than homicides in the United States, and I believe it’s like five times as many are killed by suicide than homicide in United States, maybe I’ve got those numbers reversed. And so I think it just is really trying to focus in on what are the really profound risks, I think that any dollar you spend on eliminating a risk is a dollar that you’re not spending on another risk, I think we spend a lot of money and a lot of dollars on things that don’t really give us return on investment. And I think there are things that we can get along with. So you’re talking about, you know, hey, a drone in every school that could take down an active shooter, let’s say. I think that’s insane, because the cost of putting in all of that is money that you’re not spending on counseling resources, and better teachers, you know, and social services, and all those sorts of things that are definitely going to make an impact now, and are more likely to reduce the risk of these extreme events. And I also think that, and this is where I have to be a little careful with my language, I don’t like the “never again” mentality of it. Because the thing is, is that that’s an impossibility. And I feel like- I agree with the sentiment of the never again thing, of course, we never want these horrific events to happen, right? But there is a reality of extreme events do happen, how are we best spending the resources to provide the best benefits to our society? And sometimes when it’s one of those things of like, we’re going to spend so much cash, making sure these extreme events don’t happen. My question is, well, but what if we could spend that cash on things that would produce a bigger benefit in the short term, and also would likely have a stronger benefit on these things that we’re trying to never again, in the long term. And the topic of school shootings is very, very timely right now. and obviously a huge issue in our country. And especially coming out of COVID, a massive spike in school shootings. And there’s this phenomenal foundation called the I Love You Guys foundation that does a lot with school safety. And they have, I think it’s called the standard response protocol. It’s kind of some guidance on how you talk about violence in schools, and it’s how you do drills for this. And I’m hugely opposed to active shooter drills in schools, especially those that are very simulated violence and whatnot, I’m really concerned that they normalize this violence, they make it feel inevitable. And if you have someone who is already at risk, already unstable, which if anyone remembers middle school and high school, like those are unstable times and whatnot, and if they feel inevitable, there are some really big concerns about them making it inevitable with their, and so the other thing is, there are a lot of ways you can drill that help people practice the motions without them being about a gunman, you know, you can evacuate on a fire drill, you know, you can shelter in place on an earthquake drill, you know, you can talk about mental health concerns, you could talk about violence, there’s lots of violence in our schools, that is physical violence, that’s not active shooter violence. And so hey, if we’re talking about looking out for each other, and we’re talking about minimizing fights, and we’re talking about caring for each other’s mental health, all of those things are going to directly help reduce the risk of active shooters without normalizing active shootings. And I just feel like, that that’s where I get into a little bit of the “never again” mentality is it’s like, okay, yes, I agree that we don’t want these things to happen, of course. That’s a given. That should be a given, but are the things we’re doing actually helping prevent those things again? And are they actually efficient return on investment? At the end of the day, we only have so much cash, we only have so much time and so are we spending that time and spending that cash in the right way that we’re actually producing these long term benefits for our communities with it? I’m a big fan of the I Love You Guys Foundation and their standard response protocol. I think it’s something that deserves a lot more coverage for especially thinking about, hey, we have an audience here of these students, how do we really communicate about these things in a way that we build them up, and don’t terrify them, because scaring the heck out of a bunch of middle schoolers is not going to make our world a better place. And those kids already terrified.

Seth Villegas
It definitely doesn’t seem like it would and even in the case of having something like a drone around the school or something, I’m only throwing these things out there because there’s been lots of people are trying to think of oh, what kind of technology can we use to prevent these kinds of tragedies from going on? But it sounds like you’re not necessarily thinking of safety technologies as stuff that’s just weapons or other things that are potentially dangerous, but we think about in a much broader sense. So if you’re going to kind of define what a safety technology is then, what kind of definitions would you give us?

Ryan Mayfield
Yeah, I mean, I would say that there is a place for drones in some pieces of safety technology, I don’t want to take anything away from a lot of different pieces of say, technology, my piece of safety technology that I’m really interested in, is safety communications. And so the big piece is that there are a limited number of safety people in the world. And that’s good, because we’re kind of an eccentric bunch. We don’t, we don’t need with that. But at the end of the day, each individual best knows their community, each student best knows their school. And so how do we empower individuals who see something that’s a little bit off? And I’m not just talking about, you know, the person who looks different than me whose not here. I’m talking about leaky pipes, I’m talking about slippery sidewalks. I’m talking about my friend who’s a little bit off today, and all those sorts of things. How do we enable them to build trusted connections with resources, so those can be facilities for that leaking pipe, because if that pipe doesn’t get resolved, it’s going to flood an office building and make it to a bunch of students can’t go about their work with it. If a exit door gums up, and it’s not reported and there is a fire, it’s going to make exiting much more challenging, you know, those sorts of things, how can you build those pathways to connect people who see something and know something with available resources to better escalate these and to better build this baseline resilience? And I think, this is where I can get off the rails a little bit, going into fear versus resilience. And you know, when issues emerge, the human brain triggers the fear sensors before the logic sensors. And so we are more prone to become fearful when these events come up. But there are a couple of- there’s a phenomenal book, David Ropeik’s How Risky Is It Really? which is a huge guiding light of this. One of the things we actually care about people who are afraid of flying, it’s like, well, the most risky part of flying is driving to the airport, you know, for instance. And so there are these drivers that can lead individuals to see an event as risky, or to not see an event as risky. And a lot of it comes down to things like how much choice do they have in the situation? How much control do they have on the situation? How much is there a trusted resource to resolve the situation, right? And so hey, if you see a scary event happening, and you call nine one one, and they swoop on and take care of it, right, that builds your trust in institutions that build your trust in this situation. Well, 911 is good at solving some of events, they’re not good at solving all events. So how do we better build these interplays between individuals and resources so that we can better resolve a lot of these underlying instances? And I think then there’s this bigger piece where, especially in the world of mass media and social media, we’re being constantly bombarded with fear, we’re being constantly bombarded with these events that are scary around the world. But frankly, if a bomb goes off on the other side of the world, it doesn’t really put me at risk in the short term. And so what I really want to be able to do is help people categorize events into kind of three categories, there is something I can do about it right now, number one, there might be something I can do about it in the future, number two, or I’ll never be able to do anything about it, number three, and I think most of the events that scare most of humanity right now, or at least, you know, most Americans, or those number three events, where it’s like, I’m sorry, there’s nothing you can do about it right now. And frankly, you’re probably not actually at risk from that sort of event. And so you’re venting a massive amount of anxiety and brainspace into these things, when there are pressing things in the number one and number two category, you know, that you can be actually working on now, or you can be building abilities, preparedness, or whatever it is, resources, societal connections to build in the future. And so a big piece of connecting and building these pathways to resources are that why are we not afraid of fire when we go into most commercial buildings, because there’s a fire alarm on every wall. So if you’re going to see a fire, you’re going to pull the alarm and get out of the building and if someone else sees the fire, the alarm’s gonna go off and you’ll get out of the building, great. I don’t have to worry about fire because if it comes up, I’ve got something I can do. And if someone else comes up, they’ve got something they can do. There are lots of other events like that we don’t have to find pathways to and so therefore it’s easy to worry about, like hey, if this happens, will I know about it? Will I respond properly? Will others around me respond properly? And I think that’s a big piece of resilience is having these kind of pathways. We’re not just bombarding people with like, hey, if this happens, you’re screwed. But how do we actually build pathways build trusted communication pieces so that risk specialists can better respond to risks, and people who should be focusing on getting a soccer practice, or going to school, or doing their job, or caring for their community, or whatnot can focus more of their brainspace on that, and that really is going to have a bigger impact on society than anything a risk manager could ever hope to do.

Seth Villegas
I’ve been thinking a lot with what you’re saying about how it can be really easy to dismiss communications technology as something that’s not revolutionary. So for instance, you you talk about something like 911. So you have something that’s active all the time. And I can even remember this in college, where if you have a crisis hotline or something, and it’s like, oh, that’s just the hotline, right. I’ve just, it’s so normal, that it’s- that there’ll be some sort of help that’s accessible via my phone. Whereas you also kind of gives example of, say, being in the Philippines, where you don’t necessarily have access to those sorts of things. And you’re not sure how you’re going to be able to get resources to the problems that are actually going on. Right now I- as part of BU, I work with Residence Life, which is working off in a lot and making sure you know, smoke detectors work, a lot less of the things that people need in order to just kind of stay in the building. But the more I think about like, oh, there’s all these kinds of different safety pieces that I just don’t think about, because there’s some machine there that, even if it’s just a little too hot in here, a sensor will go off and someone will come check it out. And if it is serious, then the whole building will be evacuated within a matter of minutes. And that’s kind of amazing that we have stuff like that, that we don’t even really think about because we’re at such kind of like a technological level that it almost doesn’t even register.

Ryan Mayfield
I mean, yeah, so there’s two different pieces that we can go into there. The first kind of brings it back to the dollars and time sort of thing. And so I mean, you’ve thrown out the scenario earlier of like, oh, well, what about drones that can go take down- not to pick on that. But again, every dollar you’re spending on that is one less dollar to go to maintenance, and one less dollar to go to sensors, and one less dollar to go to training and drills and all those sorts of things with it. And then the other piece you mentioned, which I’ll spend one more time about, because safety communications is you know, where I spend most of my time is this idea of well, safety communications is simple. And I think that, technologically, yes and no. I have a joke with some organizations that I work with, where if they use carrier pigeons to communicate within their organization, then just build a system around carrier pigeons. And obviously, no one uses carrier pigeons anymore, which is why I use it as a joke. But a lot of organizations think about all these highfalutin things of, ah we need apps and location data and beacons. And there are places for those. But at the end of the day, the communications pathway that works best for your community is the pathway that’s going to most effectively get them the help they need. So like if your organization is really good on carrier pigeons, don’t introduce some highfalutin app into the situation, because it’s just another barrier to entry to get access to that. So how do you come alongside how your organization already communicates? And how do you just better show them that, hey, there’s a way you can use this to get help if you need it, right? So you think about 911? Well, in the time when I was really being launched, the telephone was the revolutionary, right? No one makes a voice call anymore. But at the time, right, that was the revolutionary way to do it. There was some great scholarship, especially in the 70s and 80s, about why some people don’t use 911 so you know, we can put that on the shelf for right now. But there was a reason with this emergence there. And we are seeing a lot of really interesting stuff happening right now, some great organizations are looking at texts to 911, and you know, those different sorts of things to better align with how people communicate right now. So that’s one thing I think is really important. But you know, on the topic of how people communicate right now, I mean, I think SMS is one of the most underutilized technologies in safety communications right now. Because the simple truth is, is that everyone knows how to text, you know, my 66 year old father texts like a pro, right? And so if he saw something, even he who grew up learning on the voice telephone call, would probably still be better at texting. And so there’s a lot more-

Seth Villegas
And you’re talking about, you’re talking about texting as opposed to like apps, right?

Ryan Mayfield
Well, I think both apps and voice calls, I think it’s one of those where it’s like, like not necessarily that there’s you know, always a better thing, that you should always do text, but it’s recognizing where there’s a place for it. And so hey, maybe your organization already has an internal app that you use for productivity, and everyone already uses it, and everyone’s already familiar with it, and they’ve got to get there to get their payroll and all that stuff. Perfect. They know that platform. Use that. Maybe you’re in a community that really relies on voice, and that’s their thing that they love to pick up the phone. Great, use that. The thing is, is that for a lot of society, and just in terms of, especially in transient populations, you know, think about going to a sports event or going to an airport, something like that, right? Lots of these sports teams have their own app, lots of these airports have their own app. If I’m flying to this place once or going to one game. I’m not going to download the app. But I have texting on my phone, and I know how to use it. And I can guarantee you that 97% of the people that are passing through this airport, 97% of people who are passing through this stadium, know how to text. And so it’s about familiarity. And it’s about barrier to entry with that. We are seeing some organizations with a picking up like Slack for Safety Communications, which is great. Hey, if all of your organization communication goes through Slack, and people know, hey, when I slack my coworker, I’ll get help. Then it makes sense to add your facilities team to Slack and add your, you know, mental health resources to Slack as well, so that, hey, here’s a channel you could message with it gets the help you need from that. But especially when there isn’t one of these baseline communications, or you have a lot of transient populations, I think SMS is pretty much unbeatable. And it’s one of those things where when we’re thinking about data, which I know is something you care about a lot. I think a lot of safety teams put much of the data collection burden on the end users, on the people that are seeking out assistance, instead of putting that data collection burden on themselves as the team. So I’ll give an example of this, I have seen many web forms from safety teams of you know, hey, report a concern in the community, and it’s a 50 field form, and every single field is required, and you cannot submit your concern until that point. Well, I guarantee you a an overwhelming percentage of people turn out of that form and say, you know, this issue is really not that big of a deal, versus being able to say, hey, I have an issue, and you know, there’s a pipe leaking, or whatever it is, I am totally fine with that being the starting point, because, hey, at least you’ve registered pipe leakage. I may not know where it is, I know I may not know how bad it is. But like, I know, I have a problem. I can’t do anything until I know I have a problem, or stuff even person being like, it’s just a little drizzle, ah I’m gonna- nevermind. And then the thing is, is that there are a lot of follow up questions that can be brought in here. Where is it? How big of a deal is it? All those sorts of things. And the vast majority of people are not investigators, they are not risk professionals. As I said before, that’s totally fine. That’s a good thing. That they’re not constantly thinking about these things. Well, how do we enable them to say, Hey, I think something’s wrong. And then for investigators, you know, potentially backed up with some technology, maybe there’s some decision tree prompts, or maybe some light NLP you can spice up in there, with it, can then actually help go through, hey, is this actually an issue? Do we have a resource to respond to this? And can we get that resource out there?

Seth Villegas
I think we probably need a little bit more explanation of what you’re talking about when you say SMS as the technology. So you’ve kind of said, so far, that transient populations, SMS, and this will become clearer as we talk about it, its biggest advantage is when you don’t know anything about the sorts of people that you’re dealing with, except that they might have access to a phone and to text technology. And part of the reason for that is because when you use SMS, you can send people a sort of prompt in which they can give either a text response through a number or as you’re saying, you kind of NLP, right, so some light, you know, like a sentence or two or something like that. And the actual backend technology that hosts all of that questions is not housed on their phone, they don’t to download anything, right, they just have to send the bits of information back from each other. So I just want to make sure that didn’t get lost, because I think that’s one of the biggest advantages to SMS is the fact that the way that you can host it is only based off of having access to a number without any kind of further action needing to be taken by the person who you’re trying to get to respond.

Ryan Mayfield
Yeah, and I think that’s really important, right there, is so taking the burden off of the end user, and then putting more of that burden on the system, or the teams that are managing that system there. And so the end user, as far as I’m concerned, I know how to punch keys and maybe include a photo if I need to, or whatever it is into the system, and it just going through your standard cell phone, network connections there. But then on the back end, you can do a lot of fun things with some data analytics, if you need to, if you need to bring in location services, well, if the only way you’re actually gonna get that is if they’re on a smartphone. But if they’re on a smartphone, you can kick them a web link, they can hit that ping location services, if they need to, if it is that extreme of an event, that is also enabling people to have control over their own data. So they’re not worrying, I don’t know what all this app is harvesting off of me. But I actually have some control, which builds their own trust. You know, going back to Ropeik’s book, trust and control are major drivers of resilience there. And hey, if they’re on a flip phone, which most Americans are moving over smartphone right now, though, overseas, those numbers are lower. You’re not excluding people from the community because they don’t have an app or they can’t download an app, or they haven’t quite learned how to use an app, right? Like, I like to think of myself as fairly technologically savvy, but even sometimes I download an app, I’m like, What is this thing? That’s the last thing I want somebody thinking about when they’re trying to reach out for help. You want to make it as easy on their side as possible, and then put as many things kind of behind the curtain as possible, so that the teams that are really equipped to respond can respond with it.

Seth Villegas
One of the things that you mentioned, this point about asking the right questions, and I was just hoping you could just say a little bit more about that, right? Because that’s I think one of the things that is kind of an undercurrent of this right is like, Well, are we kind of aimed at the right thing? And then are we trying to find a solution that’s almost too technologically sophisticated for whatever we’re trying to do? Like, those seem to be the kind of two pieces that, if I was trying to ask the right questions, those are the things I would be kind of honing in on. But for you, where does your mind kind of go when you’re trying to think of like, where should I put my focus?

Ryan Mayfield
Yeah, I think there are two really big questions. One is, for a given community, what are the pressing risks? And there are some psychological elements that come into that, you know, what does leadership think is the pressing risk? What do individuals think are the pressing risks, but really like, based statistically, what is the biggest risk for this community? What are the most frequent events? What are the most impactful events? How do we impact those? And then the second thing is a hyper focus on users. How do users interact with that problem? What is, and this this kind of goes back to the timeline of events, what are the ways that individuals come to that event? Are there other events they pass through before they reach, you know, a more extreme event that we’re trying to prevent? And how do they interact with their surroundings? So do they have an established awareness of their safety team or their mental health team? Or their school counselors? Or those sorts of things? If they don’t, is that more of an organizational problem that they have? Is that a trust problem? And that goes back to some of like the scholarship in the 70s and 80s, on why people don’t call 911? There’s this great concept about the utility of recording what pros and cons people get out of reporting with that. And so yes, really focusing on who are the people we’re trying to help? And how do they interact with risks, what risks they’re presented with for it? And the thing is, is that when you ask those questions, if someone goes, Hey, great, we can put a drone in every school that would immediately stop an active shooter, my response is, well, you’re not actually knocking out the biggest risk there and you’re gonna scare the crap out of kids you’re trying to protect and normalize this idea that we need this drone so that we can do that. And you’re going to be causing a lot more negative externalities in the course of that solution, then by doing some other more proactive steps there. And I think that this is a big thing I kicked out of my defense days, which is, you know, the other side always has a vote. And I think, especially focusing on counterinsurgency as I did for a while there are multiple other sides, you know, there is, quote, unquote, the enemy, but then there is kind of a mainline population there as well, that you’re also trying to interact with and build support from and whatnot. And I think that’s really important in a technological perspective, which is to recognize that everyone has a voice in this process. And I think a lot of engineers and data scientists try to build what they think is the optimal solution. And they don’t spend enough time thinking through well, how are people going to use this? Does this align with them? And are they going to use this how I want them to use it? Or is there a better way they could use this recognizing that things can be broken? And I think a lot of the challenges that we’re seeing in the news about technological challenges or whatnot, are from organizations that didn’t think through, hey, are people going to use this as we intend? Or are there other ways that people could use this? And we are seeing some negative externalities come out from these because they weren’t thought through.

Seth Villegas
I think it would be great to get some concrete examples of the sorts of things that you’re thinking about. So you’ve kind of hinted a few different things that I know you have actually worked on. So you know, one of them is related to location data, if there’s a really big crisis, say if there’s an earthquake, a tsunami, or something like that, where it’s really important to know where people are to know if people are safe and whatnot, that’s definitely one thing. And then there’s also the having some sort of a check in process, some sort of wellness communication that comes out, you know, completely via SMS, so something that’s, you know, very easy for people to respond to, maybe it’s sent out every week, you know, every couple of weeks, depending on the various environments that it have been implemented in. But it’d be great to hear more about like, Okay, we have these kind of big ideas, what are the actual technologies that are going to be addressing these things?

Ryan Mayfield
Yeah. So let’s start at the first one, you know, location check in for events. So I’ve done some work on both app based platforms and SMS based platforms for better communicating with people in natural disasters. I think there are a couple of things that that you have to think about there. The first is how do you make sure people are aware of these events? Tornadoes don’t happen every day. Some places, they’re more frequent than others. Hurricanes don’t happen every day. And even in the midst of hurricane season, you can’t guarantee that Houston or Atlanta is going to be hit by a hurricane. And so I do think there is a lot of value to having a one stop shop for multiple issues within so if you have an organization says, Well, this is our facilities hotline, and this is our mental health hotline, and this is our security hotline. My first question is, why can’t this all be the same number because then you’re just making sure one person has to remember one number as opposed to three numbers or one app versus three apps or whatever it means. That familiarity and that trust up front, so that people know how to use it so that they’re not, you know, sitting there on 3g as cell phone networks come down trying to download an app, or trying to look up what the right number is, do they have it in their phone. In terms of location, there’s two pieces there. One is the notification process. And there’s been a lot of focus on geo fencing in the safe communications community. So hey, there’s a fire in this building, we’re going to geofence that building, we’re gonna send out a notification to everyone in that geo fence about that. And the big thing is, is as cell phone community, you know, cell phone networks, phone manufacturers give users more control over their location settings, which is a great thing, I’m a huge fan of, it reduces the effectiveness of those geo fences there. But here’s the thing, if that’s my office, or that’s my dorm, or that’s my home, I don’t care if I’m in the building, or 1000 miles away on vacation, I kind of want to know that it’s burning down. And so that’s one of those things where I think that organizations can do a lot more of the legwork ahead of time with rosters and things like that. So that, hey, I’m a university, I have a list of everyone who’s in this dorm, if it’s burning to the ground, I should probably notify everyone who’s in that dorm regardless of where they are. And then maybe I also throw geo fence on there just for good measure. Or if I’m an employer, I know everyone who is working in this tower, if there’s an earthquake there, I should probably hit everyone who works in that building, I don’t care if they’re on vacation or if they’re there so they can be notified. So that’s the outbound piece of it. And then there is the inbound piece with it. Going back to that tower example, I have an earthquake, it happens around a tower I own, I want to know about the location of who’s there so I can properly respond with resources with it. Recognizing that most people are gonna have their location services turned off, you need to connect with them an effective way that you get to the most of them and enable them to broadcast their location. And so this is one of those things where these highfalutin technology like, Oh, I’ve got this really advanced app that those super localized location services are great, most community probably hasn’t actually downloaded that app, and most of them probably have their location services turned off for it. So it may be this great technology, but it’s not actually giving you the information you need, versus say like, Hey, I can kick a push notification out to everyone, hey, if you need help hit this link. And then, you know, they can quickly say, Yep, I’m granting location services to that link, bing. And I think that’s one of those things where we can think through the user experience with this, think through the perspective of our users, and better serve them with information they really need in that moment.

Seth Villegas
We’re kind of in the scenario that you mentioned, you’re trying to get emergency information to people who live in a specific place of some kind. And the particular technology that you mentioned, was something called geo fencing. So you can correct me if I’m wrong, but geo fencing is basically taking all the cell phones in a particular place, and then kind of sending out some sort of mass message of some kind or another.

Ryan Mayfield
Yeah, geo fencing can be used, I mean, the federal government can use it to send out all levels of communication. And you can see it with Amber Alerts. There’s a number run by the state government or the federal government, there was the classic example of the false missile alert in Hawaii, a couple years back with that. And so that was through the federal government doing a full geo fence notification message. Sometimes they run an SMS, sometimes they actually run through carriers, as you’ll actually see, if you have a smartphone, Amber Alerts actually jump ahead of SMS frequently. And then a number of specific safety communications platforms will also use it on an app instead of for instance, SMS and so that it will be for users of that app who have their location services enabled, they will geo fence an area with it and it enables them to send messages just to users in that area, there are a couple different layers of geo fences that are utilized.

Seth Villegas
Okay, great. Thanks for the explanation. So one of the last things you had mentioned then that I think really sparked my interest was for most, you know, if we’re not talking about a governmental entity, right, I can just sort of do whatever it wants, one of the things that’s been happening is as technological consent has actually gotten better, it’s actually a little bit harder to get in touch with people, depending on how their settings are actually configured. So for instance, I know I do this on my own phone of please don’t just be using my location services at all times. I’d rather not be tracked all the time. But if there’s some sort of emergency going on, and I need someone to get in touch with me, right, like how the information gets me. You were basically talking about, well, how do we navigate than this problem of consent in these serious situations in which ideally, right like if your buildings on fire, or if there’s some sort of major, like a tsunami coming or something like that, right? You’d want to know as soon as possible. So how is it that you and kind of the people you’ve been working with have been navigating that sort of an issue?

Ryan Mayfield
Yeah, so the geofence issue really came up with some of the app updates that came through iOS and Android a couple of years ago, where they gave users significantly more control over their location services. And for many safety oriented organizations, they saw this as a negative because it made it harder to run geo fences because if a building is on fire, or if there is an extreme event, a hurricane, for instance, it’s harder then to geo fenced area and get that message out. I think that giving users more control of their devices is a fantastic thing, right? This idea of consent is wonderful so that we actually know what data we are sending out and what data we are bringing in for it. And I also think it brought up this important questions around what kind of information needs to get to what people? And how do we add value behind that information? And so I’m a major proponent of group based messages more so than geo fence based messages. Because if you’re a college, you have a dorm, you probably have a roster of all the students that are living in that dorm. And if my dorm is on fire, I don’t care if I’m in the dorm, or I’m at class, or I’m on vacation, I’d like to know that my stuff is on fire. I think that’s pertinent information there. I also think that it’s important to remember that there are more conventional safety measures in place. So for instance, most buildings have fire alarms. And so it’s one of those things where organizations say I need a geofence, because I need to alert everyone who’s in the building, well, you really should have a fire alarm in the building that’s notifying everyone in the building. So who really needs to know about this? And I do think that sometimes organizations don’t fully consider what information needs to go where, to whom that information is valuable. And that inherently puts a wall of mistrust between these organizations and their end users. Because the organization’s presumed that everything they think is valuable, must be valuable to the end user. But that’s not always the case. There are lots of things that risk organizations pursue that individual users don’t care about. And there are a lot of things that lead to risky events that those end users should care about there. I think this is also a big thing with 911. So many safety communication systems heavily brand themselves about being about extreme events. So, if you open up a lot of the Safety Communications apps, and you go to their websites, the first things you’ll see are active shooter buttons, or bomb threat buttons, you know, here, you use this app to report an active shooter. But when you actually talk to a risk manager, and you say, hey, if there was an active shooter on your campus or in your organization, how would you want to hear about it? The answer is almost always 911. And so I think this is one of those situations where organizations could be more proactive about thinking about consent and information flows and the value of information and how that information should be exchanged. And if they thought about it, you’d realize, hey, if like, God forbid, some extreme event happens somewhere, do you want there to be confusion about how people communicate that? And no, you don’t you want individuals to have a habit, a snap response for how to process that information. And for extreme events, it’s called 911. And there are some obvious societal challenges with metal one, there’s some mistrust of 911. And yet there are those extreme events where the answer is almost always dial 911. And so I think once you push that there is to say, okay, great, if you see a bomb dial 911, if there’s someone walking around somewhere, shooting people, dial 911, then if you’re putting this competitor up there, I think competitors are really the way to look at it, of a safety communication system be it an app or a web portal, or whatever, at an organization, users are going to look at that and say, Well, wait a second does this replace 911? So there’s confusion there. And if they know that 911 is better maintained, then it will actually degrade the performance of you know, the safety communication system. So then what really is the purpose of that safety communication system? And so if it really is keeping people informed, I think this is where the concept of the geo fence really falls apart, which is, if you really need a geo fence, it may be a law enforcement push, where it’s actually a government entity that’s overcoming these things, and they’re really pushing out, we’ve got this major issue. It may be, hey, you have an alarm system, use the alarm system. And then if there truly is something bad happening, people who are tied to that entity, to that dorm, to that campus, to that office building, it should be informed, regardless of their physical presence there. So that’s where I think you can do these things around groups. And then- so as you start to- within Safety Communications, as you sort of break down well, what risks should I be talking about? I think this is where we can actually take a lot of lessons from like marketing subscriptions, basically, hey, do you want this newsletter? Do you want to hear about daily deals? Do you want to hear- those sorts of things, because different people are going to care about different things. And by giving individuals more control over what information they want to receive, you can better tune that information to those individuals, right? So if you have someone who’s going through a lot of challenges, they may want to find out about more mental health resources. And so that’s one example. Or if you have someone who lives off campus, they may care more about traffic alerts. If you just geofence that intersection, for instance, a commuter walking through it, you know, on foot, is going to get the alert and they’re gonna say, Well, this thing is useless. I don’t want this information. And so being able to better target that information to the right people. and actually enabling and empowering people to get into the right strings information, I think is going to build much better trust between individuals and the organizations that are supposed to be protecting them and caring for them.

Seth Villegas
I think what you’re saying is really highlighting a lot of stuff for me in particular, because when I think about something like geofencing, I almost get so wrapped up in the technology that I’m just like, oh, like, Well, why wouldn’t you try to implement that? Whereas what you’re saying, you know, I work in a building where fire alarms to go off occasionally, that if you’re in the building, the fire alarm will tell you, it’s already proximal, right? Like, it’s, you can only hear things if you’re near them. So as long as the fire alarm is working, then you already have a system in place, so there’s no reason to kind of over engineer something. And it’s also not dependent upon, say if someone has smartphone access or not at that given time. And also what you’re saying about, say, with the groups for whom that’s important, again, I think that also makes a lot of sense. So if you’re going to then be pitching this to groups, how does that work, exactly? Are you trying to work specifically with organizations of some kind? Is that kind of where most Safety Communications, is it at these like kind of higher levels, because it seems that one of the important limitations of what you’re saying is that kind of roster thing, right? This type of methodology would work a lot better where you already have that information on hand.

Ryan Mayfield
So the the big thing with safety communication, is that a chatbot is not going to repair a leaking pipe. A chatbot is not going to dispatch a safety officer if there’s a situation. And so safety communications, at least for the time being, almost inherently has to be tied to an organization. I’ll take a step back, I think a lot of places over engineer Safety Communications, because they get this idea that really highfalutin risk data is going to solve the problem. And I think that that is a little bit too tech utopist. At the end of the day, most organizations already have the teams in place to solve a lot of the challenges that are there, they may need to beef those up, they may need to adapt over time and whatnot, what’s really important is connecting the person that has the issue to the team that can solve it. And so I think earlier talked about my carrier pigeon joke, you know, hey, if you’re gonna make a carrier pigeons, then just use carrier pigeons better. And that you’re saying you need to find a way to better connect your people, your community, students, employees, residents, whatever it may be, to the resources on hand, so that they can better direct their concerns to the right offices there. So it’s kind of a glorified chat system, it’s a the glorified communication system. And there have been some efforts to break these out of organizations. But I feel like they either kind of go towards vigilantism a little bit, which I think for many reasons, is very self explanatory why that’s a little concerning. I hope that’s pretty self explanatory there, I can go into it a little bit, if you like. Or they are purely informational, which I think can inspire a lot more fear than resilience. And so I have my own concerns about those as well, if they’re not using the right hand, when you just kind of broadcast them out, they kind of bring it a lot of the fear mongers or the the preppers of the world, and just kind of let them deep dive into their own world. So you really need this to be tied to an organization. So I just was able to collaborate on some research with the University of Illinois, and we really focused this on what we called Community-sourced risk systems. And we very intentionally use the term community sourced, not crowd sourced. And there’s actually some great scholars out of Korea that coined the term of community sourced information, community source safety, because it is inherently banded around this community of people. So it’s, maybe it’s not a perfectly defined community, right? Like at a university, you have those that are definitely members of the community of like professors, and students and whatnot. But then you also have contractors and visitors and alumni who are maybe quasi members of the community, that’s fine, at least you have this fairly defined piece, maybe it’s defined geographically or defined socially. And then there are somewhat defined resources available that can respond. And that’s really important because if there aren’t resources to respond to something, and someone’s reporting it, they’re ultimately just shouting into the void. And so I think there is this importance of recognizing what are the things we can respond to, and then, hey, if we are getting a lot of inbound information to a defined organization, about something that we can’t respond to, well, then maybe we need to invest in those resources. And so by having these semi defined boundaries, it actually helps individuals get access, you know, to the help they need. Individuals that desire some sort of open system want it because they recogize that there’s a fluidity to life. It’s like, well, I go to school, and I go to work and I go to, you know, my kid’s soccer practice, I go to my place of worship and whatever, like, I belong to lots of places and nowhere at once. But the thing is, is that each of those places likely has resources that can help you. And so my response to that is, well, you’re part of multiple community sourced systems, ideally, and hey, when you’re at your place of worship, the concern should be going to that place, not to work. When you’re- you’re at school, that concern should be going to that place not to work, etc. And so if you can better have individuals, where in this given moment, I am part of this community, maybe it’s a neighborhood association, maybe there’s lots of sub communities, and each of them have different levels of resourcing, we can better direct those concerns to the right places within those organizations that can then respond to it.

Seth Villegas
I think that’s also a reason to have some sort of government resources as well, right? If you need something that’s more general that reaches across different kinds of boundaries, especially the if jurisdiction’s not clear in any given instance.

Ryan Mayfield
Yeah, and I mean, I think this is one place where corporatizing the country has been deficient. There are some communities that talk a lot about how corporations basically are dependent on social welfare, right, with wages, and health care, and those sorts of things. And I would love there to be a better way to quantify that of, hey, here’s how much your employees depend on these resources there. Because that spotlight can better direct the economic power of these organizations, the voting power of their groups, right? If we know that, let’s say you’re a corporation, and you are on a county boundary, and there’s county A to the west and county B to east. If you find out that your employees are much better served in County A than in County B, that could be really important information to do that. My concern there is that you would just encourage employees all to go to county A and county B’s tax revenue would would drop. Ideally, you would use that information to beef up county B, there are so many incredible resources that exist within counties and cities and states that all these organizations depend on. When we had a more of a fixed geographic base where you know, people lived in the same town for 50 years and worked for the same company for 50 years, and those sorts of things, there was more of that defined tie. I do think like those are core threads of social fabric that are going on there. And if we can better understand how universities are dependent on city resources, and how corporations are dependent on county resources, and all those sorts of things, we should be able to better utilize that information to support those resources going forward. That piece is a little utopian, I will admit that. But I do think that’s one of the most undervalued and some of the most successful implementations of safety communication systems, especially in universities that are very well tied in with their city, I’ve seen some great information sharing, where, for instance, you have off campus students that are sending in messages to their campus dashboards and saying, Hey, I need help with this. And the best organizations are able to say, hey, we can’t help you, you’re outside of our jurisdiction, but you need to call this number or you need to text this person. Or even better, we have a relationship with them so we are going to ask them to dispatch people to your location. That’s where you’re moving towards a true holistic system of resilience there. And definitely the worst ones are those that reply and go, Hey, you’re outside of our jurisdiction, sorry, we can help you. Right? Like that is, I think that is counterproductive to both sides of the equation. And by both sides, I mean, to both the organization that is in jurisdiction, and the organization receiving the message, because it’s telling the end user that both of these are incapable of helping. I am out of their jurisdiction, but I don’t know how to get to the organization that I am in jurisdiction for. And so some of this information also that helps build better ties. And we are seeing this in corporate safety develop better, there’s been a lot of work recently, especially between large corporations and their cities, counties, states, to better collaborate with that, and to better collaboration between law enforcement agencies, social services, and these HR safety teams as well. So that they’re able to better build this holistic map of resources. And when it, when someone comes in and says, Hey, I have this issue, the corporation knows, let’s say, well, one of the biggest drivers of workplace violence is domestic violence. So we’re actually seeing some large multinational corporations stand up their own domestic violence offices internally, because they’re recognizing the huge impact on this. And so in those cases, you know, if you approach them with some domestic violence issues, they’re actually going to be built to better support you with their in-house resources. And then those teams are better informed of, okay, this is x case, that’s actually better served at the county resource or the state resource than our internal resource. And so you see those points of expertise was actually able to filter people to the right resource. And it really is that information routing, that is the most important piece. A person who needs help, get them as quickly as possible to the right resource that can help them.

Seth Villegas
I definitely like this idea of, say coordinating different kinds of community resources and then building a you’re really talking about network, right? So so people who are able to get in touch with each other and kind of reroute people to the right places. But I guess my question then is related to how do we know what information to share with what people? And if we’re kind of in all these overlapping organizational pieces in which I’m not really clear, like who needs to know what, for instance? So for instance, you mentioned something like domestic violence, right? Like, is that something my workplace needs to know about? Is that something I should just be talking to the government about? And because when we get into those kinds of issues with extraordinarily sensitive information, especially if we’re talking about, say, some sort of someone suffering a trauma, maybe that’s not information that I want to be readily sharing with everyone, even if I want to get resources for it. And so if I go to the wrong place, I might be unwittingly share information that I wouldn’t have had I known the correct place to go initially.

Ryan Mayfield
Yes, I think firewalling information, especially at initial discovery phase is essential. And I think this is one place where a lot of pop culture around police and Intelligence communities in particular has been really counterproductive. There is this image of the government in particular, as this massive all seeing, all knowing organization where all the systems are connected, and one entry into the system triggers everything. And anyone who’s ever worked in the federal government or the state government laughs at that, because to be frank, most government systems are antiquated, and don’t talk to each other and barely talk to themselves. And I think this is perhaps a bigger concern in some corporations. But then again, most people who have worked in corporate America know that getting a request through the marketing department or the HR department, or whatever it may be, is a trail of lost emails and unanswered slack messages, for instance. And so it is one of those things where I think individuals have this image that’s very different from reality. But at the same point, it is on the organization to build that trust and to indicate, and that’s where trust really comes into play here is organizations to build that trust, to emphasize anonymity to emphasize support, to emphasize the disconnection say it’s an HR support center, or whatever it is, the people working in this center do not report to the same people as those who work in payroll, or who work in employee performance reviews, and so that they can handle it separately. So there’s there’s one piece of that on the organization side, I think another big piece of it is simply the role of anonymous systems. And anonymous systems have gotten a bad rep for a long time, and there’s a really concrete reason why. Most anonymous systems historically did not allow two way communication. And most individuals are not risk professionals. They’re not investigators. And so for the longest time, anonymous says anonymous hotline, someone calls in, they report some incident, they leave out several key facts that are necessary to pursue that lead, because they’re not risk professionals, they don’t know all of the facts that are needed to pursue that lead. And they might be traumatized, they might be going through stress. And so they develop this bad rep on both sides. Because risk managers were really frustrated with Anonymous systems because they were useless, the majority of leads they actually produced was marginal. And so then individuals were frustrated with them, because they kept calling and kept calling and kept calling and nothing happened. Well, I think one of the most exciting things in modern telecommunications is that you can actually use technical systems to mask individuals for reporting while still maintaining that two way chat bridge there. And so in our study of community source risk, we actually did a deep dive into the role of anonymity in these systems. And we found organizations are far more responsive to anonymous inbound messages through one of these mass systems, because now suddenly, they have this chance to actually to ask the questions to follow up, to offer the reassurances. And we actually saw significant amounts of individuals being willing to give up their anonymity as they went through these systems, our theory and this is a theory, obviously, correlation causation. Our theory is that it’s actually an indicator of them building trust. And while looking through some of the chat logs of these anonymous systems, there’s terminology of relief. And there’s terminology of trust. And there’s terminology of confidence that comes through with individuals, of getting the help they need. And administrator asking them follow up questions. And the individuals are replying that Oh, I didn’t even realize I needed to report that. Yes, it here’s this information or whatnot. And we weren’t able to definitively show that everyone was giving up their anonymity. In many cases, they said they were willing to come into the Safety Office. So they’re willing to place a call which in case I would happen outside of the chat logs. But I believe it was a majority of the cases where there was more investigation needed, particularly, and sometimes the organization would ask like, Hey, are you willing to give us a phone call so we can talk about some more you willing to come in face to the office and give a statement with it? In many cases, individually said yep, not a problem, what’s your address or what’s the phone number? And then the chat log ends. And so we do have to make an assumption there, which is to say that they did follow through and place the phone call, or they did go in. But in the very least, they had expressed interest in actually giving up their anonymity because they were able to build this two way communication there. And so when you’re talking about consent, and when you’re also talking about data not being free, right, like data isn’t free, there’s effort that has to go into that collection. Initially, the only information the organization got was, this is someone in my organization, and they have a concern, they had two points of data there, and through those follow up messages, and maybe there were a few additional points in the concern, but through that additional follow up investigation support, they were rapidly able to uncover more leads there, and in many cases, actually find an identity and a location and these sorts of things, which really come through as important for investigating a risk.

Seth Villegas
So in this kind of a system, just so we’re clear, that sounds like it’s mainly a kind of chat interface, right, some sort of text interface, mostly?

Ryan Mayfield
Yeah, I think the text interfaces are ultimately the most important, you know, it is a person who needs help sending in communication of what they need help for. You can have like photos or videos or whatnot can sometimes be supported, you know, especially if there’s like a facilities issue of, hey, there’s a broken window, you know, being able to send that through. But at the end of the day, it’s a text with it with a description of it, and with how text familiar individuals are, you know, SMS or app chatting, or whatever it is, it naturally ties into the habit that people don’t want pick up the phone now. And so they’re sending through a text.

Seth Villegas
Alright, got it. Yeah, I think we’ve definitely kind of come full circle than in really understanding where it makes the technology powerful the potential for anonymity and two way communication. So kind of, I think one of the last things I want to ask you about, then is this potential conflict between operators and more autonomous communication systems. In part, because you mentioned earlier, say, the development of a chat bot that maybe is or is not useful, depending on how its deployed, but also these other cases where maybe you get a lot of communication at one time, you just don’t have the people to handle it. And I know, actually, in thinking about governmental entities, and even response entities, they’ve been looking to autonomous tools like chat bots, in order to increase the overall load that they can handle. Now, oftentimes, there’s also a desire in that to reduce labor, I think that that’s worth throwing out there, just because that also tends to happen when there’s the development of autonomous systems. But in theory, you could develop autonomous systems as a kind of a supplement to the kinds of people that you have, and then having operators out there filter through that information to talk to the people who need some sort of immediate assistance and whatnot. So when you’re thinking about how to develop these systems, how is it that you go about having the right response level to the kinds of problems that are going to be coming up?

Ryan Mayfield
Yeah, so chatbots provide an interesting opportunity for this, we covered a little bit in some of our past research, we actually have some follow up research right now, but specifically digging up looking into chatbots to augment safety resources. And I think the key word there is augment. A chatbot is never going to stop a leaking pipe, it’s never going to send an officer or it’s never going to be the officer on the street responding to an incident. Well, there’s a couple of threads I want to peel apart, there were chatbots may be able to have a big piece with it. First, it’s the importance of response and building trust. Second, I think it’s based off of the kind of the natural follow up questions typically come, safety and risk questions. And three, I think it’s about prioritization of messages and maximizing limited safety resources. So first, on the importance of response, one of the biggest indicators of the success as a communication system is how frequently a user gets a response and how quickly that response comes. And then I mean this makes sense, if you ask someone for help, and you never hear back from them, you’re unlikely to ask them for help again. So the faster you get a response, and the more frequently you get a response, the more trust you’re going to have in that system. That’s just one of those things where if you have a limited number of safety resources, individuals may get slower responses. And so by having a chat bot that can initially interact with them, there may be the chance to increase that response rate, and get them into the queue. So the queue brings into the next thing, right, which is that there is kind of this natural flowchart, if you will, for what sort of information is needed to respond to most risky incidents. And when you peel open a safety log, you find that there’s really like eight or ten kinds of responses that are given to inbound messages: where are you? What were they wearing? Do you need an officer? Do you- there is kind of this natural sort of piece and so especially if it can be a chatbot that is also interacting with the human and maybe helping them clarify what it is, if they can clarify it, they may be able to identify what information is missing. Ask those important follow up questions, and make sure that the situation is properly documented and interrogated, so that when it is raised to a human, more of it is done there and less has to be done by the human. And that’s really important as well, because safety departments do have fairly limited resources. And this, especially the government level, right? Budgets have been caught, you know, number of instances they have to respond to has increased, those sorts of issues. So if you can better prioritize those messages and filter those messages with some automation, you could better then direct them to the right people raise the right levels of priority, et cetera. Abig piece of the defund the police movement is focused on the need for most more social services, you know, a better connection to these resources, etc. In many places, there are some of these resources that exist already, they just require a 10 digit hotline number that no one knows instead of a 911. So hey, if there is the ability to better say, Hey, this is definitively or 95% likelihood user confirms that should be going to this department, why not kick it over to there and better do that routing, and if artificial intelligence, natural language processing, for all I care a decision tree, like people hate phone trees, and I think one thing is, is that there are some chatbots that are being built around, because making the decision tree more user friendly, which I think actually produces the most possible promise for safety with it, so it’s a little bit easier to navigate. And then this also goes back to the idea of data not being free. With too many safety organizations try to lead, you know, drop these forms out on the web, for instance of tip submitting, where they require 50 different fields to be required, well, that chatbot can quickly say, these three fields are no longer required, or, Hey, we actually need this bit of information. And there are some other things that could make it easier for individuals to interact with and provide the right information, or honestly, it’s really just like an on off form that maybe is augmented a little bit and can really make sure you’re talking to the right people.

Seth Villegas
So one of the other things that you have talked about, and I think in terms of collecting information, everything you’ve been really helpful, but one things you have written in your notes here is about it’s great to get more information, but that information can also be a burden upon the system, which is collecting it. And I feel like this is especially pertinent today in which there’s kind of efforts at total data capture. So could you talk a little bit about what you mean by data being a burden, and how the can be a hindrance in these kinds of, you know, safety situations?

Ryan Mayfield
Yeah, so many safety risk managers come out of the military or intelligence community or law enforcement. And so there is this idea that we just need more information, and we can better respond. And in these cases, I like to remind them data isn’t free. There’s a cost both to the organization and to the individuals to collecting this information. Now, the place that I care about this the most in safety communications is the risk to the user. And that really is a time burden, that really is a knowledge burden. And that’s a trust burden of why do I need to submit all this information? Who is this going to? Who am I communicating this to? Do I know all this information? And the thing is, is that when someone calls 911, to even report an extreme incident, they probably don’t know the whole thing, hi, know what I heard gunshots down my street. Which direction? I don’t know. Well, they’re still going to send officers there. And so the fact that someone is even notified, hey, there were gunshots on my street, if that’s all you have, you still want to know that information. Yes, there’s more information that’s going to come in, but you’re going to go interrogate it, you’re going to send officers you’re going to follow up, you’re gonna do all those things. Similar things can happen for leaking pipes, or mental health concerns, or sexual assaults, or all sorts of different issues with this sort of thing. Being able to simply flag there is something going on, and it deserves my attention is enough, in the beginning, there’s definitely further follow up needs to happen. Maybe NLP can help with that. But at the very least, let’s put a fork in the ground and say, something’s happening here. That’s enough to know, right now. That then brings us over to the risk to the system. Well, if you require all of these data fields to be included, you’re not gonna be a very resilient, very responsive system, because people are going to just opt out of your system. But then also, having all this information can be really problematic. I think if you look at a lot of the major data breaches we’ve talked about, in many cases, it’s organizations that are saving records that they really had no use for. And so by having this book collection, taking up petabytes of storage space, if that information isn’t actually useful to you, why are you still holding on to it? And so I think a lot of organization kind of break information down into three categories. You know, this is essential system information that I have to use. If you’re an SMS communication system, you need their phone number, full stop, right? You can’t communicate with them without the phone number, then there’s information that I might be able to use. It’s organization information. Hey, if we did truly have the setup between there’s a work instance of school incense and a place of worship instance of it, in order for each of those organizations to communicate with their users. They’re going to need to kind of know that affiliation. But people graduate from schools. And so if I still have my user roster from 10 years ago, I’m probably not actually using that information. Maybe it needs to get refiled into an alumni thing and so every five years at homecoming, it comes back in but it’s not useful just sitting there on that piece, and then information that’s never useful. Well, a full SMS communication system doesn’t know your credit card number. And so maybe there was some deal that was done through where they bought something and they pass their credit card number through it, great, run the transaction and get rid of it. Why is it still useful to maintain that sort of information there? I think this awareness of what data is useful, and for how long it’s useful, I guess, in that case, the credit card information was temporarily useful, and then it stopped being useful. Great, why do you still have it, you’re simply putting yourself at risk as an organization, and as a trusted body to do that sort of thing?

Seth Villegas
I think it’s really interesting that you bring it up that oftentimes it is that the least valuable information that’s been held, but that actually makes sense of why it would also be the most vulnerable, because you wouldn’t be taking all the necessary steps to protect it in that case, because well, you’re just not using it all the time, right? It’s just hanging out somewhere. So then it would also seem like one of the responsible things is data recycling, trashing that data and making sure that it’s actually somewhere where it can’t be recovered and whatnot, especially if there’s things that’s potentially sensitive. But also, if you’re not using it, you don’t know what information is there, you don’t know how it could be used, or how could actually make someone more vulnerable. And if I think you know, lots of large corporations today, I’m sure- I know that every time you transition from a computer system, a lot of that information from the old system gets held, but it’s not being used anymore. It’s useless, right? You kind of pored over other things. And then it’s worse than sitting a file cabinet because at least the file cabinets in a locked room. Sitting on a server that’s probably too old and is not very well protected. Because you know, who knows who’s going to need to use it when, right? And so having it be more accessible kind of seems to be the default.

Ryan Mayfield
Yeah. So in terms of safety communication, specifically, I want to touch on two examples of that. So the first is for a really app based safety communication system, which I think we’ve talked about a little bit already. In many cases, when you download a safety communications app, you’re required to go through multiple levels of user verification before you’re able to submit a tip. And for a safety organization that makes a lot of sense. I want to know who these people are. But if I just download this app, so that I can submit something, if I now have to go through seven steps, to even submit that will I ever actually get all the way to the end of it? And I think there are some things you can do in terms of saying, Hey, this is an unverified user. Hey, we don’t know who this person is, they haven’t completed their profile yet. So maybe it’s flagged, this could be more difficult as someone who’s fully gone through the system, and there’s more interrogation that’s required. And so I love systems where individuals are able to baby step their way into the system. And so it’s saying, there are ways you can better use the system by giving us more information. And ideally, you’re providing a value proposition of why that information is important, hey, if you add your phone number, we can text you as well. So that’s really important there. On the other side, in many cases, with employment systems, where you have a user roster that’s there, individuals will be getting spammed by an employee system, and they’ll say, Hey, I no longer work for this. And you can get a lot of complaints that come into a system saying, I no longer work for this. And if that user management system has been set poorly, it may be very challenging to unsubscribe that person from that setup, which will really betray trust.

Seth Villegas
Yeah, that last one, I’ve definitely have happened, right, especially depending on what the job is, it can take forever to get into the system. And then once you’re in that system doesn’t want to let you go.

Ryan Mayfield
Yeah, and it’s this idea, especially in the safety communication system, where everything we have to say is important. And the answer is, I’m sorry, everything you have to say is important to someone. But it’s not always important to everyone. And if someone leaves your organization, it’s probably no longer important. And if you make it really hard for anyone to ever get out of the system, because you say my information is always important, it’s going to make them mistrust the next organization communication system, which doesn’t help out your brethren.

Seth Villegas
I think there’s one last point, you’re talking about scraping. So social media scraping, news scraping and things like that. I see you’re kind of laughing. But I have to bring this up, in part because I think this goes along with this sort of total data capture that we’ve been hitting on so far. And especially if you are some kind of a safety organization, and you’re trying to be on top of everything that’s happening, actually, social media is one of the ways to know things that are happening. Oh, is there an earthquake? Actually, when I used to live in Southern California, one of the first things you do if you think there’s an earthquake, but you’re not sure, is actually go on Facebook to see who else is talking about potentially having an earthquake, which is really funny, because that’s definitely not something you would have had pre Facebook of not even checking the news, but just like are people posting about this right now? But also if there are other kinds of news stories that you need to know about. So that seems to be another kind of information problem that you’re trying to handle at sort of a macro level right of getting information from the environment or from people that’s like, potentially publicly available. So what sorts of things are usually looking for when you’re going about that process?

Ryan Mayfield
Yeah, so risk data is all the rage and it has been for a couple of years now. And I think there’s a lot of good that comes from risk data. There’s a number of companies that scrape social media, that scrape news media, other sorts of channels, police blotters, all of those sorts of things. And there’s a lot of really good information that comes from there. I don’t take anything away from that. And when you’re talking about travel risk management, when you’re talking about VIP protection, when you’re talking about campus protection, especially when you move into the bigger law enforcement, whatnot, there can be a lot of benefits of that. But I think it’s really important to ask yourself, Where does this information come from? And what sorts of problems are you trying to solve? And one of the big reasons why I emphasize community sourced risk is trying to better connect to defined communities because social media tends to be a little undefined, open news channels tend to be a little undefined with it, it can really connect with what are the issues in a given community there. And if you work for a large multinational corporation with hundreds of offices, or thousands of offices around the globe, it is going to be important to scrape social media, scrape news, see when all these large events are happening around the world, so you can better look out for those different offices. But if you’re trying to care for a very defined campus, or community or whatnot, what is probably more important is raising the voice of the people who actually live there who know what the issues are, and helping them better communicate. Ultimately, social media coming from all around the world is going to give a very big picture piece. But a lot of the issues that really give us resilience are small picture pieces that happen there. And if you’ll allow me a little soapbox here, there’s this crossover between resilience and fear going on right now. So the human brain there to this fantastic book, How Risky is it Really by David Ropeik, and he breaks down how fear interacts with logic and the human brain. And the human brain is wired to give a lot more credence to fear, especially in the short term than logic. And the thing is, is that the vast majority of risky events that are happening around the world are not going to hurt you. And so simply piping all of those risky events onto someone’s phone is unlikely to actually make a safer, more resilient community. Because you’re just wigging people out on these big time fears that don’t affect, now that’s splitting off being a safety manager for a fortune 500 multinational corporation, and you know, being an individual going to school or going to work wherever you may live, or work. And so the big piece I really try to focus on is that individual level of resilience, what information does that individual need? And how do we actually make them better stewards of their community, as opposed to just being fear mongers about all the crazy things that are happening 10,000 miles away?

Seth Villegas
Seems like another way to potentially have people be keyed into what’s going on around them, to have them be aware of what’s going on around them, actually, part of the issue, I imagine from a safety perspective is that people are too glued into their devices, that they’re not really paying attention to the stuff that’s immediately going on. They’re kind of deferring to external senses, external news and not using their eyes maybe falling into some sort of bystander effect. I don’t know if that’s something that you’ve seen, but sounds like that’s probable.

Ryan Mayfield
Yeah, I think the 21st century bystander effect is really challenging. And I find it concerning from a safety perspective, but I also just find it concerning from a community perspective. If individuals are more aware of what’s happening in the nation’s capital, or, you know, some overseas battleground than what’s happening in their own backyard, that’s really concerning, from just a human level, because at the end of the day, each individual is best equipped to impact their own backyard to build ties, to volunteer, to lend a hand, to help a little lady down the street, or whatever it may be. Like, if you want to make a better world, yes, there are, some people are gonna make a better world by intervening in an overseas battleground. If you’re those people, some great friends of mine are those people, and I’m so grateful for the things that they’re doing, they’re going to have a huge impact. But for the vast majority people, the biggest impact you’re going to have is on your own street. And honestly, to me, I think is a great thing, right? That’s a relief. And frankly, that person who’s helping out on that overseas background, and they also should be helping out around their street when they’re home. All of us have this duty to our own communities. We’ve been raised and taught, and we’ve worked and we’ve been helped out by our communities, and so shouldn’t we care the most about our communities? And that’s not a nationalistic thing. It’s just caring for those around us. And if we imagine a world where everyone just cared a little bit more for those around each other, well, that would link everyone. Each of us has our little radius, and then all the radiuses interlink. And, hey, that’s a really the most resilient community we have there. And yeah, there’s also that bystander effect that comes in where if you are caring most about your community, you’re going to be aware of that little lady down the street who needs some help, or the other person on the street, who maybe is how to food for the day or whatever it may be. And you’re going to find those places where you can help and then you can be able to better raise up those issues that most affect your own backyard. And so I do think there is this piece of when we get too focused on the big picture risk data News Media scraping, we lose sight of that immediate surroundings piece. And when everyone has lost sight of their immediate surroundings, that’s not good.

Seth Villegas
Definitely not. And I think that that’s a great call to action and something great to end on. But if people want to know more about your work or to follow the kinds of stuff that you’re doing, for instance, you mentioned some research that you’ve put out there, how can people follow you and where you’re researching right now?

Ryan Mayfield
Yeah, so I have a research partnership with the social computing systems lab at the University of Illinois. And so you can look into Dr. Yun Huang is the researcher there. She’s fantastic, has done a lot of great work on safety communications and interaction between police and other stakeholders. I’m up on LinkedIn if people ever want to drop on by, send me a message, shoot me a note. I’m always welcome to chat and always looking to collaborate with that. I do have a Twitter although it’s mostly about Stanford football, I’ll be very honest, three times Sports Illustrated super fan of the week when I was there, one time ESPN super fan and Stanford Daily super fan of the year. I will just kind of drop that down there for the audience that cares so much.

Seth Villegas
It sounds great. Thank you so much for your time, Ryan.

Ryan Mayfield 

Thanks, Seth.

Transcribed by https://otter.ai. Proofed by Julia Brukx.