Editor’s note: This episode contains three parts. First, Jono’s path prior to founding SameSame, a platform designed for mental health support of LGBTQI+ youth. Prior to SameSame, he has led an impressive design career in the for profit sector, and his story is inspiring for anyone who finds themselves taking unexpected turns in life.Â
Then, we move to the platform, and how it’s carefully designed to reflect the behavior, interests, and needs of young people on deeply intimate topics. Finally, the conversation turns to a broader issue: the general lack of resources when it comes to mental health. Will ChatGPT be a solution? On its own, probably not. Â
We’ll be staying on the topic of health and technology with our next episode as well, so stay tuned!
Transcript
Jim Fruchterman
Welcome to Tech Matters, a biweekly podcast about digital technology and social entrepreneurship. I’m your host, Jim Fruchterman. Over the course of this series, I’ll be talking to some amazing social change leaders about how they’re using tech to help tackle the wicked problems of the world. We’ll also learn from them about what it means to be a tech social entrepreneur, how to build a great tech team, exit strategies, the ethical use of data, finding money of course, and finally, making sure that when you’re designing software, you’re putting people first.
Our guest today began his career in the high stakes world of global advertising, designing digital experiences for giants like Coca-Cola and Guinness. But he soon realized his skills could serve a higher purpose. After spotting a dangerous lack of support for LGBTQI+ youth, Jonathan McKay founded SameSame. A digital lifeline built on the belief that even a simple text message can save a life. Today, we explore his journey from corporate design to radical social impact. Let’s dive in.
So tell me a little bit about where you’re from and what experiences, what was your career that led you eventually to starting a Tech for Good nonprofit.
Jonathan McKay [1:31]
Sure. So I’m talking to you from New York at the moment, Jim, but I am not from New York. I am from South Africa, where I spent most of my life. And I found myself after leaving high school studying graphic design. I studied graphic design without much of a plan to get into the world of social change at all. I studied graphic design because I love to draw. And it’s pretty funny to me that most of my career has been spent trying to use digital technologies to help young people make smarter choices, when I didn’t really make much of a choice about what I was studying at all. I just knew I like to draw.
I thought that a graphic designer is what you became, if that’s what you liked. And I ended up studying graphic design at the University of Pretoria. I was very lucky that the course was actually quite broad in its scope and it encompassed service design. And it really treated design as a problem solving discipline. And I think that that is something that I do almost every single day. I am using everything that I know all of my experience to solve problems and design is a specific graphic design anyway is a specific way of doing that you take elements on a page and you arrange them looking at composition, color and tone to create something that is appealing and intelligible to an audience. And I think the same applies for any digital product that you’re building, you have a set of elements, audio, video, text, set of different functionalities, and it’s your job to put them into some sort of hierarchy that’s going to make sense to your user. And that is not only going to be intelligible or make sense to them, but is actually going to be persuasive and compelling to them.
Jim [3:28]
And I would agree. I mean, I think you’re the first person who’s come out of graphic design to be the founder of a tech for good nonprofit that I’ve run into so far.
So you went through school, you acquired these skills, and then you managed to get a job?
Jonathan [3:46]
Yeah, I got a job working initially as a designer for an organization that build itself as Africa’s first convergent media company when people still use the words convergent media. And it was an organization, a company that not only offered basic graphic design services, but that also created websites and apps and also did a lot of TV branding, animation motion graphics design.
And over time, I began to specialize more and more in digital media, until the point that I became a creative director, working for organizations that were providing digital media services to corporates like Diageo, who own Guinness and Johnny Walker and Smirnoff and Unilever. Did some work for an underarm deodorant brand called Rexona in East Africa, and did some work for Coca Cola in East Africa as well. So I ended up traveling across different parts of Africa. This is around 2012. And this was happening because up until then, digital media wasn’t something that marketers were really taking seriously on the African continent.
But thanks to the explosion of mobile phone access and ownership, suddenly a brand like Guinness had a way to communicate to millions of Nigerian Guinness drinkers. At the time, it would take Guinness something like six months to roll out a new billboard campaign in Nigeria. But in the course of a few minutes, they message more than a million customers on their database, send them a text message and send a text message to members of a digital platform I helped create called Guinness VIP. That was a virtual football club for Guinness drinkers Guinness drinkers would would text in a code at the bottom of their Guinness cap. And that would allow them access to all of this exclusive football or soccer content, as you would call it, that Guinness was making available to them.
Jim [6:11]
So this is not an unfamiliar journey for many people but in the tech area, except for instead of doing digital advertising and growing up during that period, you were in this richer sort of media for major brands in Africa, kind of pioneering a lot of these techniques in those markets and being the creative that helped them go, oh, here’s a way that you could really engage your customers.
Jonathan [6:37]
Yeah, you know, one of the major advantages I feel I had, Jim, is that because of the technological constraints across the African countries I was working in, we always had to think about what an experience might be like as a single text message.
Jim [6:58]
Because the users had these inexpensive devices, yeah.
Jonathan [7:03]
Yeah, they had extremely inexpensive devices, which very often didn’t have access to the internet at all. So all they had was an SMS message or an access to your SSD or the option to call an IVR line. (An interactive voice response line). And even in instances where those phones may have had internet access, they were on 2.5G connections and the phones were still very basic. So the screen size was really small.
So I had this experience where whenever I would try and share what we were doing with someone who was living in the US or in Europe, they would be a little unimpressed, because they had already by that stage started to use the internet on their phones more, and they couldn’t understand, you know, how powerful a single SMS might be in an environment where, you know, your degree of access to technology just isn’t there. And it never struck me as a compromise either. Because if you are clever about it, you can make a single text message really powerful and really useful. And so it really forced me to think, how does every interaction with the user mean something?
Jim [8:20]
I think that’s really exciting to think of these heavy constraints on what the user experience can be, drives you to be more creative and solve those problems more effectively. But we now need to say, all right, so here you are working with major brands, helping them deliver all sorts of exciting content through highly constrained digital delivery mechanisms. How did you actually start interacting with the social good sector?
Jonathan [8:47]
So the agencies that I was working for started to do work, not just for Diageo and Guinness and Coca-Cola, but also for nonprofits. Some of the big foundations and agencies like UNICEF, Ford Foundation, UNFPA, because it wasn’t just the big corporates that were thinking about how digital technologies could allow for different kinds of relationships with their customers, the nonprofits, the NGOs were thinking, well, how do we do exactly the same thing? How do we use these channels to distribute essential information about HIV and AIDS to people who might be at risk of contracting HIV and AIDS? So I started to do work on those projects, but exactly the same kind of work I was doing for the corporates as a creative director. And I found that work much more interesting. Of course, it’s great that it has this socially useful dimension too.
So there is an altruistic factor and I could feel really good about the work that I was doing. But as a designer, I found that the problems were much more interesting because you had to think more holistically about a person’s life rather than just asking yourself, what do they need to know or think to buy another can of Coca-Cola? You have to ask yourself, “What needs to change in their lives internally and externally for them to adopt a completely new behavior, like using a condom? How on earth would you convince a man to think about getting circumcised as a way of reducing his risk of contracting HIV?” And that’s a much more interesting problem than trying to get that man to buy another bottle of Guinness.
Jim [10:49]
It’s a bigger bar to cross, yes.
Jonathan [10:53]
And so over time, I just started to gravitate to that work.
Jim [10:59]
Ah, and so was it at that same agency? I mean, I know that you worked on a very famous campaign for girls, was it at the same agency?
Jonathan [11:08]
So the organization that I worked with, it’s gone through several name changes, but at the start, it was a consulting firm called Praekelt, who started to do some nonprofit consulting work as well, enough that Praekelt established a Praekelt Foundation as a sister agency, which specialized in social change work. And I would work on projects for both the for-profit and the nonprofit entities.
Eventually, I transitioned to working exclusively on the nonprofit work, and then in that role, I started doing work for another organization called Girl Effect.
And this is an organization that was founded by Nike with the hypothesis that the best way for Nike to get the biggest return on their corporate social responsibility buck was to invest in the life of an adolescent girl. Because if you can keep a girl in school for just one more year, if you can keep her from getting pregnant for just one more year, it’s not just that her life is going to be better, her children’s lives are going to be better, and their children’s lives are going to be better.
The key to ending intergenerational poverty is to invest in adolescent girls, which is still something I believe in wholeheartedly.
Jim [12:33]
So basically, kind of a gradual thing from pure corporate, working on more and more social good things, dedicating yourselves to work for Girl Effect, and then tell us about the transition to same set.
Jonathan [12:47]
Yeah, I’m a reluctant founder in many respects. SameSame is not the first organization I founded. Very early on in my experience working as a graphic designer in South Africa, I ended up starting my own design agency. And I did that at the time because I didn’t want to work for any of the other agencies that existed. I just didn’t like the work that they were doing.
I wanted to do different kind of work. And the same is true for me now that I found it same, same. You know, I was very happy at Girl Effect. I didn’t want to start my own organization, take a pay cut, leave behind the comfort of a big NGO and start something new. I would have preferred to just carry on. But I felt compelled to start same, same, because I think that digital technologies hold a unique potential to support LGBTQI+ young people. And I just didn’t see anyone doing a good enough job in the places where I think they have the greatest potential to do good.
There are lots of great organizations doing work in the U.S. and in Europe, and I just didn’t see that happening in Africa. When I looked at many of the grassroots and community-based LGBTQI plus organizations working across Africa, it didn’t seem to me that they were investing a lot in digital technologies at all, and quite rightly so in many instances, because they are having to make very hard choices about what they do. And often they’re thinking, how do I find a bed for someone to sleep in tonight? How do we find a way of compensating for the fact that this young person has lost their tuition? How do we desperately petition the government to prevent punitive law from being passed by parliament? They’re not thinking, what is our social media content calendar looking like? And what is our digital product strategy looking like? And why do I say that digital technologies have a special role to play for this audience? I say that because face-to-face support is often not available to them. LGBTQI+ young people live in very hostile environments where what they’re going through is a secret that in some instances, they must continue to keep, they must continue to keep the secret. They cannot talk to their family. They cannot talk to their friends. They often feel that they cannot talk to their teachers, and in many instances, they are correct. They must not talk to those people. Those people cannot provide them with the unbiased support that they need or the safe space that they need.
Digital technologies, however, that’s perfect. Here’s a way that you can anonymously interact with a person or a service without revealing too much about yourself and get the support that you need wherever you are, even if there isn’t an LGBTQI+ organization in your neighborhood or that one gay uncle that everyone knows about. You could be in a remote rural area and still get the information you need. I know that digital technologies offer this potential for young people because I experienced it myself as a young person.
When I was growing up in South Africa struggling with my sexuality, I knew that there was a gay bookstore that I could go to. I knew that there was a clinic that I could go to. I knew about these places because I page through the classified section of our local newspaper, the Pretorian News, and right at the back with all of the very kind of skeezy ads, they’d have ads for this gay bookstore or this clinic. So I knew that they existed, but there was no way that I was going to go to that clinic or that bookstore. Not only was I afraid of being seen, but there was something about the idea of stepping through the doorway of any of those places, which felt impossible to me. I felt in my head like I would be somehow coated with an invisible ink that would be sprayed from the door frame that would mark me out as “gay, gay, gay” forever. But I could go online to forums and blogs and create pseudonyms and leave comments and questions and interact with people without that same fear that I was marked. It wasn’t a commitment. I was just playing around. If I’d gone to the gay bookstore, that would be a point of no return. I’d definitely be gay. But as long as I was online, there was a little bit of distance. And so I had really profound experience when I was young that left me with the conviction that digital technologies do have this role to play for.
Jim [18:03]
Sort of a gray zone for exploration as opposed to a black, white, you know, you’re jumping off the cliff as it were.
Jonathan [18:10]
Yeah, yeah, yeah, yeah, exactly. So I had that personal conviction.
And then professionally, I just didn’t see anyone was taking advantage of that. And there were so many projects I was working on where it didn’t make sense to do anything specific for the LGBTQI+ community, often because we were working on youth empowerment programs, where there was a partnership with the government or a network of schools, and to try and open up a conversation with those partners to say, hey, what could we be doing to include sexual and gender minorities in more meaningful ways that could potentially imperil the whole setup. Or at least it felt that way to me, you know, we’re trying to get a Girl Effect, we’re trying to convince people that, you know, their girls are valuable enough to send to school, it didn’t feel appropriate to also start having a conversation about what being a girl is, or means.
Jim [19:09]
Wow, so you see this gap, you see that the existing structures and programs aren’t meeting this need. You don’t really want to make the leap into starting your own organization, but you were compelled to do so because you felt like someone had to do it. You looked around and said, no one else is doing it. I guess it’s me.
Jonathan [19:29]
Yeah. I don’t think that’s going to change ever, Jim. I’m always going to have that feeling that I’m doing this work until someone else is doing it better than I am, and then I’m going to try and find out how I can help them and contribute to what they’re doing. I want the work to be done. That’s what I care about.
That’s one of the very frustrating things I found about starting up an organization because in the early stages, I was much more concerned about the work and not the organization. But so much of the money that I was able to secure at the start depended on me building up an organization before I even knew whether or not the work was worth doing, before we had any data about whether or not we could build something. There seem to be so few because everyone wants to know like, where is your anti-money laundering policy? Where is your procurement policy? Who are your board members? Before they even want to know, do you have any indication that the service of yours that you’re building is going to work?
Jim [20:39]
This is crazy, and it’s one of the reasons why I push people to say, you should start as a fiscally sponsored project of a large organization that has the audit, has the policies, does all that, so that you can work on product market fit and not anything else. But for example, in Africa, they’re just, there is not as robust a physical sponsorship infrastructure as there is, say, in a country like the United States.
And who knows? Maybe fiscal sponsorship really is just a creature of U.S. charity law, right? So, well, let’s pivot to, you know, what is same, same, what did you guys do? What was the breakthrough? What made you feel like it was working?
Jonathan [21:22]
The idea for Same Same first came to me in 2012. I was working on a project called Young Africa Live in South Africa. It was a project that was using mobile technologies to provide young South Africans about information about HIV and AIDS. And very cleverly, the editorial direction was not to talk about HIV and AIDS at all, because these are not topics that are interesting to young people, but to talk to them about love, sex, and relationships in the time of HIV and AIDS. And in the course of that project, we put together lots of stories, including stories about gender and sexuality. Some of the stories about gender and sexuality performed very well from a traffic perspective. We had plenty of people reading those stories, but the response from users was really toxic. Comments along the lines of, you’re going to hell, I don’t want to point that I realized I had built a platform that was not safe for LGBTQ plus people. And I put a deck together. And the idea presented in the deck was that there should be a digital safe space for young people. I imagine it as a mobi site that encompassed a forum and different kinds of FAQs and live chats with experts that would cover everything that an LGBT white plus young person needed to know to navigate adolescence and young adulthood successfully. And that was the original idea in 2012.
But I didn’t do anything with that idea until 2021. When a friend of mine let me know about this accelerator program that was looking for ideas for the ways that WhatsApp in particular could make the world a better place. And they had remembered my idea from 2012 and they nudged me in the ribs virtually and said, you need to dust off that deck because I think you need to submit it to this accelerator program. The original idea was for a range of different services across a range of different digital channels. When I started to think about what I would submit to this accelerator program, I thought WhatsApp is the perfect channel. So it’s on everyone’s phone, it’s encrypted, it’s cheap, ding, ding, ding, it’s popular. Young people are already, and it’s not just that it’s already on their phone, they’re using it every day. And a starting point for Same Same is that we want to be where young people are. And then the second change from the original idea, which was broader in its scope of services, was to focus on mental health. And that is because I thought to myself, we cannot replace the work of the excellent organizations on the ground in South Africa, in Zimbabwe, in Zambia, in Kenya, in Nigeria, we want to complement what they are doing, we want to be able to direct users to those services, and we’ll never be on the ground ourselves. Like, we recognize that at some point, your online journey may need an offline, you know, continuation, a few offline steps. That’s not our job. So we were thinking, what can we do that we could measure with a high degree of rigor to know that we were making young queer people’s lives better, that would allow them to take full advantage of the services that are being offered by these other organizations.
And human agency seemed to be what became a central part about the theory of change, and specifically how mental health mental well being affects human agency. If you are struggling to get out of bed in the morning, you are not going to have the energy to go to that LGBTQI plus friendly clinic to get under prep. If you are so anxious, because you are gender nonconforming, and you’re facing discrimination every day that you can’t even muster up the courage to go to a job interview, well, then you’re never going to get that job. And so we decided that mental health would be the focus.
Jim [25:49]
Okay, so you’ve identified the opportunity. You didn’t really want to do it, but you felt compelled to do it.
You applied for an accelerator. One assumes that you got the money and then you had to do it. Okay, so you had a prototype. You’re scaling up.
What is the product?
Jonathan [26:07]
Yeah, so before we designed the product, we went about looking for existing mental health interventions that had been designed for the LGBTQI+ community, where there was good data to show that the intervention was effective, that it was improving mental well being. And we found an intervention called a firm that had been developed by a group of academics from the US and Canada that had already been tested across multiple countries, not in Africa, and not the kind of lower end of the income spectrum. But at least here was some evidence that you know, this is an intervention which is which can work across different contexts. And there were multiple peer reviewed papers behind it.
And this intervention was an eight session cognitive behavioral therapy course that brought young people together in groups once a week over the course of eight weeks to learn the coping skills and kind of reframing skills that cognitive behavioral therapy imparts. And we determined that we could adapt this content and deliver it through WhatsApp in a conversation in an entirely automated conversation. But we were building for scale from the start. So we considered whether or not first we could have a person deliver this material to a user through WhatsApp. And then we just ran into the problem of that’s never going to be scalable if we’re starting with a person we want to be we want to start by testing out the most scalable version of what this intervention could be. And we also considered whether or not we might use WhatsApp groups for a while, because the original intervention was designed for groups. And for reasons relating to safety and security of our users, we opted not to do that. We opted to do it like a one on one chat with a bot. This is pre AI times. So it was a kind of decision tree style or menu based style WhatsApp experience where we use menus and buttons and keywords to help users make their way through this course. I will sometimes describe it as something like Duolingo for mental health through WhatsApp.
Jim [28:39]
Wow, is it that kind of chat therapy?
Jonathan [28:42]
No, it’s not exactly. It’s closer to education in many respects. Okay, the building blocks of the CBT course that we provide users include things like a goal setting module. And that goal setting module helps explain what good goals look like, why it’s important to have goals, and then helps the user understand this is relating to what good goals look like, what a specific goal entails, how you might assess whether or not a goal is realistic or not.
Is it feasible? So in this, at the start of the module, assume the user doesn’t have a lot of goals is just kind of freewheeling through life, and doesn’t have the tools to, to set achievable measurable goals either. And by the end of that 20 minute module, they have a very specific goal with a timeframe that is achievable. That is, you know, that that is something they can do to make their life better. That’s one module. There is another module focused on social networks. This module asks users to reflect on the people around them in their lives, and to ask who they might think of as peers, who they might think of as supporters, who they might think of as role models, and to understand that we need all of those kinds of people in our life that they play different roles. It asks you to reflect on where you think there might be gaps. And then you put together a plan to plug those gaps, or to address issues with your existing social network. And that’s because we know having a good support of social network is really closely correlated with good mental health, with mental well being. So in the course of that 20 minute module, you leave with a plan to improve your social network.
And whether or not you act on that plan, because it’s going to be a smaller number of users who go, you know, who adopt a new behavior, at least you’re, you know, you’re equipped to know, like, what does a good relationship look like? We have a series of these different modules, each of which has a concrete output for the user, which collectively provide them with a set of tools they can use to cope with stressful experiences and improve their mental well being.
Jim [31:13]
And your assumption is, based on the research, is that if we can get more people, more kids, through these modules, we’re gonna be improving their mental health. So that’s your… so completion is a proxy for mental health improvement.
Jonathan [31:26]
Yes. That’s how things started.
I think we have a more nuanced approach to measuring what we’re doing now, and that is because, and this will be unsurprising to anyone who’s built any digital product and looked at user funnels, it’s very challenging to get a person to go through 11 different modules, which is where the current version of our intervention starts. Also, each individual module imparts valuable knowledge and provides you with skills.
Now, we’re not looking at completion of the entire course as the only thing we care about. We care about the completion of each of the individual modules because each of them individually is beneficial to a user, and we’re trying to get better at understanding the dosage situation. What is the optimal amount of modules? Which modules? In what combination? Over what period of time?
Jim [32:27]
So people don’t have to go through them in this rigid sequence. They don’t necessarily build on each other. They can stand apart.
Jonathan [32:34]
And there’s a lot of interest in the world of mental health in what are called single session interventions, Jim. And this is, these are interventions that are designed to be completed in a single session. And that’s based on the reality with this, this area of research has emerged in response to the reality that for many people, their first session of therapy is their last session of therapy. I was fortunate enough to attend the single session symposium, I think it was called in Chicago, a conference earlier this year, hosted by Dr. Jessica Schleider and her lab for scalable mental health. And during one of the sessions, they flashed up a slide which covered what she calls the single set mindset. You have to believe that change can happen in an instant, that that change can last. And that people already have everything that they need to change.
And the third point, you know, people already have what they need to change, really resonated with me because it, it reminds you that you are, you’re acting in service of your users. And this goes back to what we were talking about earlier, Jim, about making sure that you’re delivering something meaningful to them. You know, they are, they have their own hopes and desires, which sometimes align with yours. And sometimes they don’t, you need to be thinking very diligently about what you’re asking them to do, and whether or not it’s actually making them feel, better or different right now. And to do that, you have to work with the resources they already have at their disposal, not imagine, you know, that something in their life might change tomorrow, because you have no control over, over any of that.
So we, our bot is live in South Africa and in Zimbabwe. To date, over 360,000 people have engaged with our bots. Now looking at users so far, about 71% of users are testing positive for symptoms of depression. And that we find for users who do complete the CBT course, over 80% of them show improvements in their mental health for the latest version of our bot in Zimbabwe. And just under half of those who show improvements, show improvements that are significant enough to make a change to their clinical diagnosis. So we don’t do any diagnoses ourselves, but we use the standardized clinical tool, the patient health questionnaire nine, that’s used to measure depression on our bot. And we serve that to users before they begin the course and after the course. And if you if your score changes five points or more on that scale, then that’s enough to take you from a diagnosis of severe depression, to moderate depression or from mild depression or from moderate to mild or from mild depression to no depression at all. So we have 86% of users who are showing positive shifts on that scale. And then just under half of those users are showing shift of nine points or more. And we’re also sending that that second PHQ survey out two weeks after the first one is is set.
Jonathan [36:15]
So we don’t have data yet about whether or not those improvements persist three months, six months, a year later, that is something that we are gathering. But we have that data to show that using this standardized clinical tool, our users mental health is improving.
Jim [36:34]
And of course, the great thing about a digital course is you can be doing all these measurements. I mean, you’re measuring, how do you like the course?
You’re also measuring whether they’re less depressed. I’m amazed that you get that many people to cooperate in completing these questionnaires, extensive questionnaires.
Jonathan [36:50]
Yeah, it’s really interesting, Jim, because here’s an example of one of the tests that we did, which would seem to suggest that young people are not going to provide you with information like this. So in the beginning, our onboarding process included having the bot ask the user for a nickname, and we were very clear this doesn’t have to be your real name. And we found that when we ran a version of that onboarding experience that didn’t ask that question, it improved onboarding completion rates by 17%. Wow. So, you know, that our immediate conclusion was, do not ask the users any questions at all. And this is because they are living in environments that are very hostile to their LGBTQI+ identities. And we thought, you know, they just don’t want to, even if it’s a nickname, they just don’t want to share anything like that. And so we haven’t been doing that.
But recently, we just started running into problems with segmentation, because we were just like, well, we don’t know whether or not this user is trans. We don’t know if they’re a lesbian. We don’t know how old or young they are. So we want to start collecting more demographic information. And we just we ran, again, a little test, you know, one version of our onboarding experience where we ask no demographic information and then another version where we do some basic AB testing. And we found that the majority of users are very happy to tell us their gender. They’re very happy to tell us their sexuality. They don’t want to give us a nickname. But they are happy to tell us their gender.
And when it comes to the patient health questionnaire, the way that it’s framed when we ask it is like, before we start the course, you need to know where you know how you’re doing right now so that you know if you’re improving, you know, you know. And the more you tell us about how you’re feeling right now, the better equipped we are to help you. So I think there are also ways to this is about providing value for a user. It’s like, what’s the value in me telling you what my name is? I don’t care about that. Oh, but I maybe there is some value in me sharing how I’m feeling right now if that’s going to help me track my emotions over time.
Yeah, so thanks to a grant from the Patrick J. McGovern Foundation we built a reinforcement learning system which allows us to run multi-armed bandit style testing where we have different versions of our bot operating at all times. And have users randomly assigned to one of those arms and then have this learning algorithm that is monitoring constantly which of these arms are performing better and then diverting users to the best performing arm. But it never stops testing some of those other arms unless there’s some real problem with them because our assumption is that conditions may change over time. You know that there may be a good reason that video is not doing well now in this, you know, in one context that we could launch the bot tomorrow, you know, or three months from now in a different context and things might be different. And maybe that’s because there’s a new data bundle that makes video cheaper.
Jim [40:05]
This is the thing that flies in the face of traditional, you do a five-year randomized control trial, and it’s like, do you know how fast technology changes? And you just described multiple reasons about why, and a change to the pricing could change whether or not you’re going down this path.
You know, the political environment could rapidly shift in their country, and suddenly things that worked great two years ago aren’t feasible anymore, they just don’t, and so you just have to be that flexible. And I think, this is, I think, one of the most exciting stories about what you guys do is not only are you doing all this testing, but you’re doing it all of the time.
Jonathan [40:45]
Yeah, there are a number of different tests that we are running all the time. And one of the things which we’ve seen, Jim, since LLMs became more ubiquitous, is that the performance of our rules-based bot has degraded. It’s not doing as well as it was in the past.
And our assumption is because people now are through Instagram and through WhatsApp, they’re interacting with metas, LLMs. And so we have to respond if we’re going to stay relevant.
Jim [41:18]
Yeah, that’s because in some ways, your mental health bot is competing with the general, the available chat bots that are out there, and if you are seen as being behind, then people are going to be less engaged with what you’re doing.
Jonathan [41:35]
Yeah, that’s such an important point, Jim, because when I think about who we’re competing with, we’re not competing with other mental health services, because young people are not going to those mental health services, you know, let’s assume they exist, which many times is not true, they do not exist or not accessible. But even when they are accessible and available in a specific location, young people are not going there, because it’s boring, because they don’t want to because they don’t understand why they would.
So we are competing with TikTok for young people’s time and attention, not another service that they’re not using.
Jim [42:14]
So one last area I wanted to talk to you about is this larger issue of mental health globally, right? If you take a step back, the services by and large aren’t available, the humans to deliver the services are by and large not being trained, I mean. And so there’s this giant gap between the amount of mental health burden on societies that people are carrying around all the time and solving that.
You obviously are a pioneer in using chatbots, AI, to deliver mental health services. And you’re considering expanding. I mean, do you actually think that three to five years now we’re gonna have really effective AI counseling in mental health?
Jonathan [43:09]
No, I think it’s fair to say that we are in the midst of a global mental health crisis, and that the situation is particularly acute for young people. And that the solutions that we have today are simply not up to the task of addressing this crisis. There are roughly nine trained mental health counselors, mental health workers for every 100,000 people globally. And across the African continent, that drops to 1.4. And there have at times been, for example, only one psychiatrist in the whole of Malawi. So there are not nearly enough people. And while the global mental health movement has done a lot to expand access by training lay workers through what’s called task shifting, that is not happening nearly quick enough to meet the need.
So I firmly believe that the only way to meet the need is to introduce digital technologies to the mental health ecosystem in a more meaningful way. And to look at step care models that use digital tools, not only at the start of a person’s journey, but particularly at the start of a person’s journey where they are being assessed, where I think digital technologies have a large role to play, where there is some triage where you’re trying to understand this urgency of a person’s need and making sure that they get to a human being quickly, if a human being is what they need. But an important caveat for me is I don’t think that a human being is necessarily the best option for everyone at all times. We have many users who have told us that they prefer talking to our chatbot, because they’re not afraid of judgment in a way that they would be if they were talking to a human being, that’s not going to work for everyone. And we certainly do have users who are asking to speak to a human being. But I take it as a given that there are there are people out there, there is a significant group of people who would prefer a digital service.
And so I think that there are ways that digital technologies can provide people with skills that can prevent the onset of mental health challenges. There are ways that digital technologies can help us assess mental health challenges and direct people to the resources that they need at all. And then I think that there are ways that digital technologies can support the whole of a stepcare model by also doing things like collecting data along every step of the process, using technology to manage case records more effectively to train lay workers to provide lay workers with tools to improve their skills a lot of the time. So I expect there to be an ecosystem in which digital technologies and AI play a significant role.
Common Sense Media recently published a report into young people’s use of AI companions in the US. And one of the things that I found really interesting in that report is that young people are using AI companions to practice having conversations with their friends, and that they still value the conversations with their friends more highly than they do the AI companions.
So I am not there are certainly people who are more predisposed to form unhealthy attachments to AI companions. And there are ways that the makers of many AI companions are unethically designing them to produce dependence. So I think that, you know, there are plenty of bad actors, including most of the big, you know, LLM producers now who are not putting in place efficient safeguards. But I think that the risk can be managed. And that the alternative to do nothing is unconscionable, considering how dire the situation is.
Jim [47:36]
Well, I think that’s a very, you know, nuanced and articulated sort of approach of we have to act, the commercial players are not acting right now in the interest of most people, surprise. And it seems so unusual to me that at this moment where mental health burden is surging, the need, the unmet need just keeps increasing that organizations in this field are struggling to find the funding to be able to do this innovation in a way that actually benefits the people we serve.
Well, Jono, I think this has been an incredible conversation. We’ve covered such a wide array of topics. And I think that this question of mental health is so important. Thank you very much for joining us and for sharing your knowledge with our listeners.
—
A huge thank you to Jonathan for sharing his journey and the vital work being done at SameSame. It’s a powerful reminder that sometimes the most important use of technology is simply creating a safe, anonymous space for someone to be themselves.
If you enjoyed this episode, please follow, rate and review the podcast on your favorite platform and share with someone who cares about mental health, LGBTQI plus rights or human centered design.
If this conversation resonated with you, I’d love for you to check out my new book, Technology for Good, How Nonprofit Leaders Are Using Software and Data to Solve Our Most Pressing Social Problems from MIT Press. You can find links and more details at Fruchterman.org.
I also want to acknowledge the support of the generous donors who make this podcast possible, especially Okta for Good. I’m your host, Jim Fruchterman. Thanks for listening.