The Social Innovation Podcast spoke to Hannah Thinyane, a Principal Research Fellow at the United Nations University Institute in Macau. Hannah leads the Migrant Tech Research Project. Since 2016 she has led a multi-disciplinary team, innovating and inventing ICTs to support proactive and consistent screening of workers in situations of labour exploitation and human trafficking.
She has over 15 years of academic and practical experience in the area of mobile computing, ICT for development, and human-computer interaction. During this time, she has undertaken applied computing research, conceptualizing, designing, developing and rolling out systems for underserved areas in Africa and Southeast Asia. Through this work, she has made strong connections with mobile operators, local and national governments, inter-governmental organizations, industry and civil society organizations.
She has authored more than 90 peer-reviewed publications based on her research, publishing in both academic and policy circles. Hannah’s work has been presented at / showcased by: UN Special Rapporteur on Human Trafficking (2019); UN Business and Human Rights Forum (2019); World Justice Forum (2019); and at INTERPOL’s 2018 Global Conference on Human Trafficking and Migrant Smuggling. In 2019 she was invited to speak to the Police Advisers to UN Member States (Nov 2019. Australian Mission to the UN, NY) on using technology to support human trafficking victim identification by law enforcement. She has also served as a panelist at the Global Fund to End Modern Slavery’s webinar on using tech to assess working conditions in supply chains (June 2020); Code 8.7: Using Computational Science and AI to End Modern Slavery (Feb 2019. UN Headquarters, NY); and the Thomson Reuters Anti-Slavery Summit (Aug 2017. Hong Kong).
Read the transcript of our conversation below.
Technology Gives One the Ability To Generate Evidence That Can Be Used To Inform Decisions
Michael Waitze 0:00
Hi, this is Michael Waitze and welcome back to the Social Innovation Podcast. Today I’m joined by Hannah Thinyane, a Principal Research Fellow at the United Nations University Institute in Macau. I’m impressed actually, Anna, how are you doing today?
Can you give us a little bit of your background for context, and if we’re gonna laugh through this thing, that’s even better.
Hannah Thinyane 0:42
That’s good. Because I tend to laugh a little bit otherwise. Okay. My name is Hannah Thinyane and I am at the United Nations University in Macau. And mostly, I’m from Australia. I did my PhD in computer science in Adelaide. And then I decided I wanted to do some exploring. So I went to I moved to South Africa, and did a postdoc, and at Rhodes University in a small town called in Grahamstown, South Africa, looking at ICT for development. And then maybe four years ago, and I moved to Macau to take up this position. I’m really looking at ICT for sustainable development.
Michael Waitze 1:28
So did you study computer science as an undergrad as well?
Hannah Thinyane 1:32
But I did math and computer science as an undergrad, as well as a Bachelor of Arts in international study.
Michael Waitze 1:39
So in the team where you work, is it filled with people…I’m sure it’s filled with people with PhDs, but is it filled with particularly in your department…Is it filled with people with PhDs in computer science?
Hannah Thinyane 1:49
No. So the team is quite small. On my team, I’m the only one with a PhD. And we all come from different backgrounds, someone from economics and public health, international affairs, I’m not sure where everyone else got the degrees.
Michael Waitze 2:07
It is just interesting to me that, given the context of what you’re doing now, maybe you can explain that in a second as well, just like an overall general, then I’ll kind of jump into the main topics, but how do you get from a PhD in computer science, which in and of itself is kind of interesting and awesome, into what you’re doing now?
Hannah Thinyane 2:27
Oh, I’ve always been interested in helping people in general. And after I got my PhD, I was like, I didn’t know. I didn’t necessarily want to go into just industry, I want to see if I can use these skills, and to help people. And that’s why I really went into ICT for development. And so if you think of this, this field of ICT development, and some people would be very much tech-focused, you know, so, and they’d be focused on seeing how can we use technology? How can we use mobile phones? How can we use this? And then you might need a lot of computer scientists around? Sure. But if you take a problem focus and say, What is the problem out there that we’re trying to address? And then just a real understanding of, of the needs and the context and all of this stuff? And from that we’ve been kind of built the other way up to say, Okay, well, then how can technology play a role to solve that problem? So in that case, you need people who are good at interviewing people,who understand economics, who understands all of these other fields. So we can take the understanding of what you need to do, or what role technology could play.
Michael Waitze 3:52
When you were growing up, was it kind of predetermined like, did you figure out early on in life that you were really good at math and science? Was it something that you always excelled at? In other words, was that path into a PhD in computer science, I wouldn’t say it was predetermined, but it was it obvious to you at an early age?
Hannah Thinyane 4:13
The reason I did math and computer science at university in Adelaide was because that degree was really open. Like only half of your classes had to come from math or computer science. And I figured, well, then I could do stuff from other fields. And like, I did philosophy and anthropology, you know, these kind of things just to see, well, what do I actually want to do? And yeah, and then I realized, hey, actually, I’m quite good at computer science and a major, a double major in computer science, because I realized I didn’t actually like math.
Michael Waitze 4:49
I think that’s so cool. So tell me exactly what do you do? In other words, what is the thing you’re trying to figure out at the United Nations University Institute in Macau?
Hannah Thinyane 5:00
What my team is trying to do is we’re trying to understand how we can use technology to support migrant workers in vulnerable situations. Now we understand that all work situations, you can map onto this continuum. This is some, like I can give you a link to who said it, she says that all work environment, and can be put on a continuum from decent work at one end to forced labor at the other. And so we’re all kind of on this continuum somewhere and you kind of move from one end to the other through various different types of exploitative circumstances and may be violations of criminal law or an illegal and that kind of pushes you towards the forced labor side. So with this continuum in mind, we were we went to understand how we can support migrant workers on any point, but probably towards like the forced labor side, and to use technology to change their work environment, and change the conditions of exploitation that they face.
Michael Waitze 6:11
So is there are a specific or group of specific situations, common situations or industries in which workers get exploited? In other words, Is there like a top-five places, not places, but industries where you can just go there and know if we can find the right and in some cases, the wrong people that we know that they’re going to be exploited workers there? If you’re going on that continuum, you’re talking about from decent work to forced labor?
Hannah Thinyane 6:35
Yeah. And so, unfortunately, what happens is people in any precarious work situation. And so if you think of agriculture, anything like picking fruit, vegetables, getting sick and raising animals, and anything, that’s what we would say, is a low paid job. So maybe also construction, often domestic work. Anywhere, there’s a hard to reach environment to assess the labor conditions, that’s when we really see lots of exploitation because people know they can get away with it, because no one’s going to be watching.
Michael Waitze 7:18
Right? And is it region-specific? In other words, when you look around the world, is there plenty of migrant workers in the United States in California that are picking grapes, or, you know, throughout the US that are doing work that I would consider not necessarily forced labor, but it fits into that category? So is it region-specific? Or is it just everywhere in the world?
Hannah Thinyane 7:36
It’s everywhere around the world. If you look at that, they say that there’s quite a lot more in Asia Pacific, because that’s where there are lots of people who, if I’m honest, they may be exploited, and in factories to provide goods for the rest of the world. Yeah. But definitely in America, Australia, Europe, there are lots of people and who are exploited as well.
Michael Waitze 8:03
So exactly what are you doing specifically? Are you studying it? Are you going out and trying to fix these problems?
Hannah Thinyane 8:10
Yeah, so we started our work in 2016 2017. Um, and I started actually in Thailand. And because we had connections through one of our partners, then Mekong Club over in Thailand. And so we had a series of working groups, where we met with what I call front line responders, they, anybody who really has the mandate to assess the working conditions of workers, and to help them and get support if they want them. So this is like the police and the labor inspectors, and different intergovernmental organizations like IOM, and we work with NGOs, and we even got to meet with survivors and shelters as well. So we met with all these guys to say, Okay, so in and victim identification, what are the phases that people go through to, to make a case basically, right, and the different groups that we took into and first of all, there’s an initial screening, which happens out in the field, and in front of, well, essentially, anybody really, after you got done this initial screening, and if you realize that, firstly, the worker went out. And secondly, there are grounds to stick out. And then you open up a case. And then there’s this whole case management thing. If we’re talking about labor, exploitation and forced labor, these are legal definitions. And so the government definitely at this stage, and then it goes through the process of going to the court. So when I looked at this, I really thought like the case management and the part where the government that’s really set in stone, there’s not much that we can do there. But it really looked like that first stage of initial screening was a place where we make an impact that we took to all these different types…
Michael Waitze 10:12
I just feel like so I feel like I see this every day in my everyday life. And I’ll tell you why. First of all, between the border of Myanmar, so Burma and Thailand, there’s a lot of sort of worker migration, maybe not so much during COVID. But this is something that’s as well known. And a lot of those workers end up in multiple industries, whether it’s, you know, in the fields like you said, for agriculture, but also in construction. And I see it actually in my neighborhood, where every time there’s a construction project, in my neighborhood, they put up these little shanties, for lack of a better term, I don’t know what the conditions are like inside, but on the outside, it doesn’t look super. Now, I’m not sure that these people are forced into this labor. But there has to be a way. And I know that I know that sort of organizations like the Baan Dek Foundation, are trying to work with some of the construction companies to make sure that the children that actually live there, right, because if you have parents that are working, something has to happen to those kids, I can see some of that happening. But now let’s talk about this. How do you use technology? And you mentioned blockchain earlier, we can get to that if you’d like. But how do you use technology? And what role can it play in just making sure that you know things are better properly monitored, people, aren’t forced into labor that contracts are enforced? Or that people can still maintain their passport and their ident? Like all these little things that happened to people that are in forced labor? what role can technology play and feel free to be as detailed as possible? Because I’m really curious, I see it in my day to day life, right?
Hannah Thinyane 11:40
Yeah. So um, that’s exactly the question we went out asking. So now that we said, okay, we want to look at that initial screening in the field, see if people need help. And we say, Well, how could we use technology and the groups that we talked to came up with mad suggestions, like, sure some of them, but we should use their phone to make a portable lie detector. So we can see if people are lying, whether it’s the boss season playing, and when you ask him about working conditions, if the pain people are withholding the documents, but even the workers themselves to see if they are actually just a bit grumpy with their boss and trying to get him in trouble. And other things, we’ll talk to that, you know, we should have like a whole big brother thing, have all these surveillance cameras, and all this kind of stuff. And I really wasn’t quite interested. And but some people said, Look, we have this problem, that we can’t actually talk to people in the field that freely first of all, people might be listening. Second of all, we all speak different languages. And thirdly, we don’t necessarily know what the key indicators of exploitation are. So we know that exploitation, practices of exploitation change over time, because it’s like a game of cat and mouse, right? If I’m an exploiter, and I know you guys now looking for this, but I’m not going to quite do that, I’m going to do something else. And then it takes like the good guys I have doesn’t quit happening here. But the good guy, and a little while to figure out and the practices have changed. And so we developed this system called Apprise that allows us to make lists of questions. And the questions are based on the most recent and most common indicators of exploitation that people in the field have seen. And then we translate these questions into different languages. And then we have them in audio format on a phone, so we can just ask, and the different workers in their own language, about the practices of work that they experienced. And so the idea is that the system has a bit of extra effect, an expert system is what we call it. So there’s a knowledge base in the background that knows how to combine the actors to say how vulnerable their situation is. And then that informs the work of themself as well as the frontline responder who’s there to say, hey, it looks like it’s a very vulnerable situation, and that this worker’s in and they’ve talked about withholding wages or debt bondage, or ID retention and all of these specific factors so that the frontline responder can use that information to inform their next steps. So if the frontline responder is, say, a Ministry of Labor person, we were working with the PIPO centers port in port out inspection centers in fishing in Thailand, if it’s like a labor inspector or a Navy Inspector, they have a mandate that says for example, if the person is a child, and he sees the child on a fishing vessel, he must be removed. And other than that, they don’t necessarily remove him that there are other the system gives advice of what the next steps can be based on both the title as well as the responses from the workers and staff.
Michael Waitze 14:57
Is there a way to use technology to disintermediate from, like at the hiring process? In other words, if you know if you have some idea, right, because you may never really know, but if you have some idea who the bad actors are just based on, you know, years of data and years of experience, you know, and you said debt bondage, right, you kind of kind of glossed over it. But you should probably define that for people that may not understand what that mean. I’m guessing that that means that the bigger you actually pay to get that job, and then your repayment rates are so egregious, or usurious, that you can never get yourself out of debt. So having a job has no benefit to you, but you still have to work to pay back that loan. Is there any way? I mean, and that’s a terrible situation, right? I mean, imagine if, you know, for me to do my job. So I have to pay, you know, $100,000 to do it, and it’s 18% interest monthly kind of thing. It’s, it’s impossible to take yourself out of it, right?
Hannah Thinyane 15:52
Yeah. Or if maybe your family had to sell their farm back home and whatever else that they had as a means of getting income, because there was this promise of a great job, right? And now your family didn’t have a way of earning income and you having to pay a whole heap of interest on a loan, and then you’re not actually being paid yourself. At the rate that you agree to? That little bit.
Michael Waitze 16:16
Yeah. So this is where I really want to know, in other words, at the point of hiring, I have almost everybody has a mobile phone, whether it’s prepaid or postpaid, it that also makes a difference to right. But everybody’s kind of carrying around this, you know, supercomputer in their pocket? And how can you I really want to understand this because I want to help out in some way, right? How can you use that tech to know where they are? Because most of these phones have GPS in them to prevent and fishing is a terrible example of this. I mean, it’s a great example, but a terrible situation, where once you get on a boat, you can’t even escape. Right? So how do you use tech, and make sure that whoever hires that person, brings them on the boat allows them to have their phone that they have the right connectivity, like maybe all these boats should have Wi Fi on them, or some kind of satellite connectivity. So you can always know where they are kind of thing. I just want to really want to drill down a little bit and figure out how detect can actually be used to help you when you mentioned blockchain to Is there any way that that plays a part? I have some ideas, but I want to know what you’re seeing.
Hannah Thinyane 17:23
So it’s a really tricky question. Yeah.
Michael Waitze 17:27
Yeah, there’s no direct answer. But I just want to have that sort of back and forth to understand what other people besides I am thinking.
Hannah Thinyane 17:35
Yeah, one of the big problems that I have is, “who do you trust?” in this situation. And with all of the intermediary, the recruiters and all of these kind of things in the middle, then how do you know who the big bad guy is that you’re actually taking loans from? That’s kind of that’s part of the problem, that also these people who are deciding to do jobs, just because I wouldn’t want to do it. And I think it’s exploitative. Does that mean that they shouldn’t do it? You know, who’s the person who should be making that decision? I think the agreements, and they have limited job opportunities. So maybe the thing is that they need to have more job opportunities at home. And it’s kind of it’s really tough. And then we like to paint this the bad guy and the good guy, but there is no, no person. Well, I don’t think there is a person who’s always bad, and there’s no person who’s always good. You know, say what if like, in the interviews with workers, one of the things that that we’re always told is, what if the worker is lying? What if they’re just making false claims about the conditions and just trying to get their boss in trouble? So just because a worker calls for help, do we believe them? And how do we get evidence about what actually is happening? These are all difficult questions.
Michael Waitze 19:05
It’s really tricky. I wish we could kind of map this out on a whiteboard, right? So you still just take one migrant worker back in his or her hometown? Yeah, the best solution for that person, I think, you know, what I’m spitballing a little bit would be to have them have an opportunity to work and earn money in their hometown. So you don’t have to be a migrant worker. But let’s put that aside as a possibility, right? Because, and that’s a problem that technology probably can’t solve, at least not in the context of the conversation that we’re having. Yeah, but let’s say that they see some kind of an I don’t know the way this works, but some kind of advertising, right, like get on a fishing boat in Thailand and earn enough money to send your daughter to school, whatever it is, right? It’s it’s alluring. So there’s an agent in the middle. One of the things that I learned in corporate life and one of the really terrible things about corporate life is that your boss also has a boss. Yeah, that boss also has a boss and that boss has reported to the you know, Board of Directors. Then they have shareholders and fiduciary responsibilities. And you’re right, like, your boss may be super mean to you, but you don’t know what’s happening to him above him. Right? You don’t know his boss or her boss is putting a ton of pressure on them, which then feeds down to you. So you may get on a fishing boat with some super nice people. But the guy who owns that boat, or the gal who owns that boat is just egregious and terrible. And I just wonder, like, you know what I mean, right? And I just wonder, like, how does tech solve that? Where do you really just need to put somebody on every boat, which doesn’t scale? as kind of like the TSA police man like you do on an airplane? You know what I mean? Like when you get on a plane? So you look around, you’re like, Okay, who’s the lady who with the gun that’s really going to protect all of us if something bad happens, because she’s probably on the flight, but you don’t know. Right?
Hannah Thinyane 20:50
Yeah. So what I think in Thailand, they’ve done all of these reforms, and they’ve got so many levels in place now. And in response to that yellow flag that EU raised about fishing conditions a few years ago, and that kind of a thing is “are they motivated to find people and do stuff about the cases?” And when, when they, if they do identify someone to be in a vulnerable situation. And there’s like, part of the thing I like about technology, I’ve always been interested in the ability for you to generate evidence that can be used to inform decisions. And so transparency and accountability for me, I really the key things that we can get from Tech. And that we can say, I say where the price the system we generated. And we developed when when you do an interview, it creates a log of all of the responses. And that gets kind of uploaded into my organization’s account, if I’m the frontline responder, and then we can go through it. And so if my boss wants to see, they can see who I have been interviewing. And they could follow up and say, Hey, and actually this interview that you did today, it said that the person was in a highly vulnerable situation. And there was fines of whatever, withholding wages documentary attention. And and they could say to me, what did you do about it? And so there can be that pressure on me. But then that, really, that there needs to be that person who is pushing for accountability involved, whether it’s my boss in my made-up organization, or whether it’s like a, I don’t know if this is an intergovernmental organization, is it a group or a panel of NGOs, or somebody has to be able to access data and be able to follow up and stay and demand accountability to action.
Michael Waitze 23:02
Right, so one of the things that I think about in the I don’t even want to call blockchain, but sort of the distributed ledger technologies, this is, again, let’s take one migrant worker, who before they leave their hometown, you know, they have a phone that’s DLT enabled, and they get some kind of communication that’s been recorded on a ledger, right. And everybody knows about it. It’s it’s, it’s available. Yeah. So everybody understands what that conversation was. And then they sign a contract, let’s say, and because the contract sits on a distributed ledger technology, if it’s an open DLT, that everybody can see that as well. It’s freely available. It’s been approved, right? And so then, you know, when that person signed that contract, you know, the details of that contract, you know, and obviously, if it sits on Ethereum, or some other automatically, happening, contract platform, you know as soon as that contract gets executed, that some kind of payment has to get made. And you can then protect against the sort of what did you call it?
Hannah Thinyane 24:03
The debt bondage?
Michael Waitze 24:05
Yeah. So you can remove the the ability for them to do debt bondage, at least if everything has to go through the chain, there can be a record of it, it can’t change, you can’t amend it is one of the things that I’ve heard is that, you know, a worker will sign a contract. And then the owner of that contract will just change those terms and file it away somewhere. When the when the worker complains, just go look, here’s what you signed.
Hannah Thinyane 24:27
Yeah, so there’s two things that that and the first thing is, it becomes a bit big brother, because, um, I might not take a loan out from my employer, but I might take it out from somebody who has money in my hometown. And then so who else goes on to my big DLT? You know, and who has access to all of this information? And if we’re saying that other people should have this DLP would you be happy if your contract to secure and lending history goes on to this publicly accessible thing, so that kind of brings up one privacy side. And, and the second thing I was gonna say was about the contracts. And that’s a really cool idea. This one, I’m not necessarily a blockchain person, but there are two paths where I can see what blockchain and DLT playing in, in the work that we do. So Diginex has this product called eMin. And that’s really for contracts to stop contract substitution. And so when a worker gets a contract, the thing is scanned and placed on a DLT. And then everybody can have access to this thing. And so you can make sure that that there’s no contact substitution there, I think that’s a really useful thing, the only other time that I could see a ledger could play is in that accountability that I was mentioning beforehand. And that maybe if, if I have, at the end of all of my interviews, it says there were five cases, and that we identified as vulnerable. And those cases should then go into the ledger of some type. And I don’t know, it shouldn’t be a public ledger, but there should be people who have accountability and oversight, and who can demand, um, to see, like, some evidence of a case being followed up. I’m not sure how that would work. But surely, if you have additional people being identified as vulnerable, maybe you’ll see an increase in the number of people trying to seek shelter. You know, maybe they’ll be more legal cases or something, and opened up, maybe we’ll just have more entries into a case management system. And but there needs to be some evidence that we can check to say, okay, people are doing their job. It’s not just getting to this point. And then nothing happening.
Michael Waitze 26:56
Yeah, I mean, you make a really good point, right? In other words, I don’t want my employment contract on a publicly available blockchain. It’s nobody’s business, right. But on the other hand, I don’t feel like I was being exploited as much. And I say that with a clear conscience as much but not not that I wasn’t exploited, but it’s not the same. It’s not the same category, right. And, yeah, there’s got to be some balance and you make again, you make a really good point between somebody’s privacy because you don’t know that you’re right, they may have borrowed money off chain from somebody that has no way of being recorded. So they could be in debt bondage. And something that’s completely unrelated actually, to the the agent that recruited them, the business owner that employs them, it could have been a cousin or an uncle or somebody else in their town, it says, if you want to travel from here to the place of your employment as a migrant worker, you’re gonna have to pay me some egregious fees to do that. I can see that happening as well, that happens outside the scope of this particular employment. You’re right. So it’s a it’s kind of I don’t want to say intractable, right, because I don’t think anything is unsolvable. But it’s very, it’s much more complex than it looks on the surface in the sense that you don’t always know who’s involved. And you’re right, identifying the bad guy in quotes, can sometimes be really hard.
Hannah Thinyane 28:11
Yeah, we’re doing some work on another project where we’re looking at recruitment and use of technology in recruitment to overcome some of the power imbalances, and all of this stuff. And, and we were I was interviewing someone from Nepal and the guy was saying, and I was talking about how recruitment agencies always rip people off and if they think that there’s a role for technology to support them, so they don’t get ripped off. And the guy just laughed, and he’s like you don’t know what life is like, you know, here, it’s your brother, his sister, it’s your husband or your wife who will sell you. And because they want some money, if a recruitment agency charges me a bit more than they’re supposed to, that’s nothing in comparison. And so it just, it just stuck with me because the guy he’s like, you don’t understand what life is like actually outside of your little bubble.
Michael Waitze 29:08
That’s also a really good point, right? In other words, we are transferring our cultural norms, right? And say, yeah, and saying My sister is never going to lend me $1,000 and charge me 18% interest and yet, maybe, right because we don’t understand it’s not that she would I’m not saying that that’s I’m not associating that with any particular cultural place. But what that gentleman was saying to you is you can’t generalize from your own life experience about what’s happening here. Yeah, can’t even create a policy or even do research based on the way you think our lives are.
Hannah Thinyane 29:41
Yeah, and that’s culturally acceptable. I don’t I personally don’t think it’s acceptable but in that culture, it’s it’s like that’s what people do. You take a taxi, someone rip you off, you go to the shop someone or if you offer everyone struggling, and say it’s just it’s It was a really good reminder about how I don’t know, things are different in different countries.
Michael Waitze 30:09
Yeah, and in different cultures, and are you fortunate enough, at least before COVID, to be able to travel to some of the locations that you’re studying, and interview those people face to face? Or is everything you do essentially remote.
Hannah Thinyane 30:21
Right now, to see it, it’s all remote. Previously, and yeah, we were in Thailand every week, every month or so. That was really, it was really such a privilege to be able to go so many times and really see how things are changing. And it’s actually how we would maybe I’ll give you a bit of background. So we were doing working with a few NGOs, and, and a few of the PIPO, port in port out inspection centers. So we need some of the teams in the in the centers and the NGOs. And when we would visit around from from center to center, we would see how and just by taking part in different interviews, practices of exploitation, we’re changing it with time. And so the first month, and some of the practices you would see in one port, and then you’d go to the next port, and by the next time you go there, they were doing the same thing. So people are learning how to get around inspec tors. And so I had this idea. Last year, beginning of last year, I think it was I remember exactly when and but this is and if you think of epidemiology is the study of disease, and he study how diseases spread. Yeah. And they change over time. And and so we said it from a theoretical kind of point of view, saying what if we say human trafficking, and labor exploitation like a disease? And what if we try and borrow some of the approaches that we people use in public health, and see how we can apply them to labor exploitation. So there’s this approach called Sentinel surveillance, where you have surveillance in different sentinels by staring at a sentinel fight, it’s just like a place where you’d think you’d have a greater than average chance of coming across, and someone with the disease, in this case, labor exploitation, and they that would be maybe a PIPO or the NGOs, places where there are people who are in vulnerable position. And then you equip them with testing facilities, to be able to rapidly diagnose the disease, then escape, exploitation. And then you collect information from these sites. And you bring it together and see if you can monitor. But now all of this kind of language, we can add brain automatically thing. So COVID, checking all those things. And but I really think there’s this stuff we can learn in labor exploitation as well, do you have things change? Because then if you know how things are changing, you know, the current practices, then you can look at prevention, and you can look at better protection. So if you know how people are currently exploiting others, you can train your frontline responders what to look for now, rather than what were the the patterns of exploitation, say if he is the guy when he would be initially changed?
Michael Waitze 33:24
Yeah, I mean, it’s really interesting, right? Because if you look at it from an epidemiologist standpoint, and you think about, you know, disease spread, but also disease mutation, which is what you’re talking about, when you say you go to one point like this, you have another port. It’s the same but slightly different. It’s hard to attack it. Is there a way to apply artificial intelligence and machine learning to things? And is there a way to gather the right kind of data? Because you made a really good point earlier. You can go to a factory owner and say, and I think, you know, Leanne, and I joked about this a little bit, too, you can go into a factory and just say, Are you exploiting your workers? And none of them are gonna say, you got me? Yeah. And if the workers are standing around when the factory owner is standing there, you can’t just walk in and say, Is anybody here being exploited then like six people raise their hand? Right?
Hannah Thinyane 34:14
Yeah. So that’s why I developed Apprise to be able to help frontline responders to interview people and to collect the responses. They the interviews are “Yes and No Questions” like Are you under 18? Do you have access to ID documents? Right? If you don’t have access, can you access them and some of the lists, like 40 questions, some are 10 questions. But you get these lovely, yes, no binary answers. And which, as a computer scientist, very nice to be able to analyze.
Michael Waitze 34:45
There’s no gray between zero and one, right?
Hannah Thinyane 34:47
Yeah, so we actually have Yes, No Maybe. I don’t know if that’s the thing is, so I have a One, Zero and Minus One. Kind of long term, I think it would be really great to do some kind of pattern detection. To do this, you need a lot of data. And so it with artificial intelligence, you need a training set of data that is very different from your testing set. So you’d have to keep on using it in the field to be able to kind of get this set to then look at seeing if you can detect anything meaningful.
Michael Waitze 35:23
Yeah, and it’s again, another great point, you don’t want to be too big brotherish about this. But you do want to be able to present people, right. So you don’t wanna install cameras everywhere. You don’t want sensors everywhere. But you can actually do surveys, I think, and you’re right, if you have the right amount of data, you can get this reverse indication. So you don’t have to be physically in every factory to have a sensor on every fishing boat. If you gather enough data, you can just figure out, that’s a red flag. And even if it gets a little bit, you can still say that’s still cancer kind of thing. But takes time.
Hannah Thinyane 35:54
Yeah. Yeah, it takes time and what we do at Apprise that if you start seeing something you can add in question. And that kind of goes a bit deeper into that whenever you go see, you know, and so then the next time the frontline responders log on to their phone, it’ll download the new question and all the new translations of that question, so that you can then get a little bit more information, which will help you narrow down what’s happening.
Michael Waitze 36:21
That sounds really cool. So what is your day to day like, if you don’t mind me asking? In other words, you just constantly trying to improve the Apprise system? Are you? Are you studying all this stuff all the time? And let’s remove again, COVID? I know you said you’re in Thailand once a month. So that’s kind of cool. What’s the day to day like? Yeah, sitting in your office, when you’re in Macau? What are you trying to figure out?
Hannah Thinyane 36:43
It really depends. I have a lot of different things I’m doing. I know. If I think of Apprise only, and quite a lot of fit, if I think of last year, not this year, because this is just
Michael Waitze 36:55
yeah, it’s just crazy.
Hannah Thinyane 36:57
And yeah, we’ll ignore 2020.
Yeah, so in 2019, I, what I was doing is I was running a series of evaluations of Apprise of different people, as well as me going to Thailand, I also have a project coordinator in Thailand. And she would go and meet with the different frontline responders in the different parts of Thailand. It’s better for her, because whenever I go anywhere I stick out like a sore thumb andI don’t speak Thai…
So she could go and check to everyone she follows up with them all on line and all the time. And say we would do evaluations with different people of new features, for example, say one of my big things was I wanted to check the question were accurate and the translations were using the right kind of phrasing and the right terminology. These translations I really had to get right. Because Yeah, not only does I have to be language specific, but it also has to be specific to the kind of colloquial phrasing that you’d use if you were in the target demographic. If I was someone who would work in a factory processing fish, it would be a very different person to a professional translator, and then having the right terminology for things like contracts and payments and things like that to understand how people in that kind of context would refer to things. So quite a lot was looking at a series of different evaluation, then i am the programmer as well, by doing any changes and tweaks to the system in response to whatever findings we have an academic so we have to do research and the writing of of these findings up for researchers, as well as policy making.
Michael Waitze 39:02
No wonder why you are so busy.
Hannah Thinyane 39:06
And I say yes to too many things.
Michael Waitze 39:09
I know I know that feeling directly, but what I’m working on is not nearly as important as what you’re working on. Look, I really appreciate your time today. This was really interesting. I’m gonna let you go. But I want to thank you very much.
Hannah Thinyane 39:23
Thank you. It’s been fun.
Michael Waitze 39:24
I’m glad you enjoyed it.
Hannah Thinyane 39:27
Transcribed by https://otter.ai