He Helped Train ChatGPT. It Was Traumatizing. – With Richard Mathenge

Richard Mathenge was part of a team of contractors in Nairobi, Kenya who trained OpenAI's GPT models. He did so as a team lead at Sama, an AI training company that partnered on the project. In this episode of Big Technology Podcast, Mathenge tells the story of his experience. During the training, he was routinely subjected to sexually explicit material, offered insufficient counseling, and his team members were paid, in some cases, just $1 per hour. Listen for an in-depth look at how these models are trained, and for a look at the human side of Reinforcement Learning with Human Feedback. --- Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice. For weekly updates on the show, sign up for the pod newsletter on LinkedIn: https://www.linkedin.com/newsletters/6901970121829801984/ Questions? Feedback? Write to: [email protected] ---- OpenAI's response: We engaged Sama as part of our ongoing work to create safer AI systems and prevent harmful outputs. We take the mental health of our employees and our contractors very seriously. One of the reasons we first engaged Sama was because of their commitment to good practices. Our previous understanding was that wellness programs and 1:1 counseling were offered, workers could opt out of any work without penalization, exposure to explicit content would have a limit, and sensitive information would be handled by workers who were specifically trained to do so. Upon learning of Sama worker conditions in February of 2021 we immediately sought to find out more information from Sama. Sama simultaneously informed us that they were exiting the content moderation space all together. OpenAI paid Sama $12.50 / hour. We tried to obtain more information about worker compensation from Sama but they never provided us with hard numbers. Sama did provide us with a study they conducted across other companies that do content moderation in that region and shared Sama’s wages were 2-3x the competition. Learn more about your ad choices. Visit megaphone.fm/adchoices

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Alex Kantrowitz?

  LinkedIn Presents.

• 0:14Alex Kantrowitz?

  Welcome to Big Technology Podcast, a show for cool-headed, nuanced conversation of the tech world and beyond. Richard Matenga is our guest today, and he's here to share the story of how large language models like OpenAI's GPT model get trained, because he and a team of colleagues in Africa actually did it. Matenga, who is based in Kenya, is a former team lead at Sama, a company that's trained AI models on behalf of companies like OpenAI, and his is a story you really need to hear. And fair warning, there are parts of it that just are not pretty. While at Sama, Matenga and his team reviewed and rated texts based off of quality and explicitness to help make the product that you and I use

• 0:55Alex Kantrowitz?

  today a pleasant experience. This involved routine exposure to some extremely awful texts, which Matenga will describe here. In this conversation, he talks about the human side of reinforcement learning with human feedback, which is the key advance that helps make these models so impressive. Too often, the human element of that advance is left out. And so today, Matenga is going to share it. He's going to share his story of how he and his colleagues, needing work in a pandemic-erect economy, found themselves at the ugly center of one of this generation's biggest technological advances. My conversation with Richard Matenga, coming up, right after this. This episode is brought to you by Genesis. What does it actually take to put agentic

• 1:38Alex Kantrowitz?

  AI to work across an entire enterprise? I recently spoke with Genesis chairman and CEO Tony Bates about why orchestration is emerging as a competitive advantage in customer experience, and what companies need to get right in the AI era. Then, I sat down with Adam Mitchell, head of enterprise business solutions at Voya Financial, to go inside a large-scale agentic CX transformation, including the business decisions, trade-offs, and lessons along the way. You can watch both conversations now on my YouTube channel today. It's never been more challenging to get Frontier compute. How do you know where to look, who to trust, or how much to pay? Akash is an open cloud that provides easy access, flexibility, and peace of mind.

• 2:23Alex Kantrowitz?

  That's how companies like Razor were able to cut their inference cost by 15 times. Visit akash.network/alex for a free consultation. That's akash.network/alex for a free consultation.

• 2:38Alex Kantrowitz?

  Richard, welcome to the show.

• 2:40Speaker 01

  Thank you very much, Alex.

• 2:41Alex Kantrowitz?

  So to start, Richard, where in the world am I finding you right now?

• 2:45Speaker 01

  I am a resident of Nairobi, in the outskirts of a country called Kenya in Africa, a place by the name Mbakasi, and that's where you can find me.

• 2:56Alex Kantrowitz?

  And talk a little bit about the beginning of your career before you started working in tech projects or as you started working in tech projects. How did you get there?

• 3:05Speaker 01

  Right before I started doing tech, I was engaged with insurance. I was a salesperson doing insurance. Right after insurance, I moved to customer service. Largely, that's where my experience, most of my, or vast of my experience lies in. I did this in an organization called Technobrain, right before moving to SAMA.

• 3:36Alex Kantrowitz?

  And what is SAMA?

• 3:38Speaker 01

  SAMA is an organization that deals with artificial intelligence. It has its headquarters in the US. One of its branches is right here in Nairobi, where it has been operating for the last couple of years.

• 4:00Alex Kantrowitz?

  So what drew you to SAMA as a company? What made you want to work there?

• 4:05Speaker 01

  Well, right, as I was doing customer service, you will definitely want to, as an ordinary human being, you will want to grow in terms of career, just like any other individual. It was a dream just to move to a different organization so that you can grow career-wise.

• 4:30Speaker 01

  As many will have done, you made an application and you will pray that your application will receive positive response. And hence, that's the reason why I tried my level best to make an application. And that's how I found myself in Sama.

• 4:52Alex Kantrowitz?

  And it sounds like it was also an opportunity to work on something pretty cool, some cutting-edge new technology to dig into artificial intelligence and help train it. Did you know initially that that was what you were going to be doing?

• 5:07Speaker 01

  Absolutely. Just like any other individuals, as I said it earlier, it was something that you were looking forward to. It was an euphoria, you know?

• 5:22Alex Kantrowitz?

  Right. So you start at Sama in July 2021. What does the work that you're doing look like in the beginning?

• 5:30Speaker 01

  In July 2021, we were actually engaged in LiDAR projects. They were doing LiDAR as well as, you know, content characterization. And so for me, I was actually alongside my friends, we were introduced to LiDAR projects. And LiDAR project, the LiDAR project that we were introduced to, it was really engaging and it was really interesting at first. But right as time went by, you know, interrelationship between the employees as well as the management was becoming unwanted.

• 6:09Speaker 01

  And we, we found ourselves in the, in, in, in the wrong foot, you know, of the organization, where they didn't want anything to do with us. And so basically, that's how, that's how they relate with people.

• 6:26Alex Kantrowitz?

  I want to get into the management and employer relationship first, but I'd also like to understand a little bit more about the nature of your work before we jump in there. So just to begin with the LiDAR project, is that like helping to train self-driving cars? Or can you talk a little bit about what that was?

• 6:43Speaker 01

  Absolutely, that's that's basically what just as just as you put it, it is rightly what we were engaged in.

• 6:53Alex Kantrowitz?

  Oh, cool.

• 6:53Speaker 01

  It is self.

• 6:54Alex Kantrowitz?

  Yeah, what does that work look like? What are you doing when you're when you're doing that training?

• 6:58Speaker 01

  The training was all about annotating annotating cars, something which you know will be seen in future. Basically annotating cars, moving objects and all so on and so forth.

• 7:18Alex Kantrowitz?

  So you have like a view of the road and you're saying, okay, this is a car, this is a bicycle, just identifying, labeling what's going on.

• 7:26Speaker 01

  Absolutely, absolutely. That's that's.

• 7:29Alex Kantrowitz?

  Interesting.

• 7:29Speaker 01

  Really, yes.

• 7:31Alex Kantrowitz?

  And so how do you move from there to start starting to train text models or chatbots?

• 7:39Speaker 01

  So right, right after LIDA, just as I said, we we we had a friction with the management. The management didn't want anything to do with us. They they made, you know, commands and and instructions which didn't go well with some of some of my colleagues and I. And we found ourselves on the wrong side of the divide. And so the next thing we we saw, we were actually laid off.

• 8:17Speaker 01

  They have this arrangement, some has this arrangement where if you if, you know, if your contract comes to an end, then they put they place you

• 8:32Speaker 01

  on the bench, which means you. You are not engaged for a period of time up until when they find it fit uh for you to get back to the organization. Uh should they find another, should they get another project? And so I was lucky enough um to be called uh for another project uh which is uh the so-called chat G P T. Where you uh are given a text, then you benchmark uh the text uh according to several benchmarks which uh are provided for. So that's how I found myself leading a team of eleven, sorry, ten individuals.

• 9:15Alex Kantrowitz?

  And so, did they tell you that this was going to be for a chatbot or just for the G P T model that OpenAI was using? And what did it look like when you were actually working to train these models?

• 9:29Speaker 01

  Firstly, we went through a training. The training was about large texts or uh small texts. Uh sorry, short texts. Uh and those texts, you're given several benchmarks. I I I think it was six or seven. When when you read the text, you are you are actually instructed on picking the closest benchmark which will um based on the instructions, sorry, the specifications given by the client. So for instance, if you you're given a um a a text, um several benchmarks, uh I believe

• 10:08Speaker 01

  one of them was uh illegal, erotic, non-erotic. I I seem not to remember the the others. But if erotic is uh if erotic is um

• 10:23Speaker 01

  the most severe, uh the most severe in terms of the specifications given by the client, then you choose that that benchmark.

• 10:32Alex Kantrowitz?

  So it it sounds like it you basically went screen by screen, looking at different pieces of text that this model would generate, and then you sort of had to rate it on a safety level?

• 10:43Speaker 01

  Absolutely, absolutely. So that's the case. You you you pick the most um the closest uh benchmark uh to save on um to save on the text that you uh that you have read.

• 11:00Speaker 01

  So that's basically about ChatGPT.

• 11:03Alex Kantrowitz?

  And Richard, I mean, we've all seen what this these models can produce now. But when you were training them, did you have a sense as to like what the product was that you were working on? Or what what did you think that this product? I mean, LIDAR is obvious, right? Like selecting images for self-driving cars. Okay, simple, got it. Rating text based off of these different guidelines? I mean, what do you imagine this product was going to look like?

• 11:30Speaker 01

  Looking at where it is right now.

• 11:33Speaker 01

  First of all, I could not, we could not even tell. Remember, Alex, when when we when we started engaging ourselves with ChatGPT? It was during the COVID era. We were just coming, we're just going through the COVID. COVID period. And most of the most of the individuals that that were engaged in that project were looking forward to something which, you know, will make ends meet. We were going through a recession. Being in that situation, anything that came our way,

• 12:16Speaker 01

  we will just take it, you know? And that's how we found ourselves there. We engage ourselves in in assignments or obligations, which, you know, were very traumatizing, texts which will make you think and overthink and rethink. And a lot of things were were going through our minds. We didn't imagine we will be here. We didn't imagine that we will get we will get to this point where we read things. Some of some of the text messages are not even in a position to. Disclosed to you just because of the nature of of the text.

• 13:00Alex Kantrowitz?

  Yeah, I'm getting the sense that the test, the text that they threw at you was was pretty awful. You're sitting at these screens, you're getting, you know, basically batch after batch of text trying to rate it. And I imagine this is what's trying to make these models ready for prime time, ready for people to be able to use it and not see things that will horrify them. But in order to do that, you need to go through this training. So talk a little bit about the the type of text that you could see, not to the point that you feel like disgusted by what you're saying, but tell us as much as you can.

• 13:39Speaker 01

  The the the closest and the most traumatic text was one that, you know, described a situation which is unlikely to happen, unimaginable, so so to say, in our society, a situation where a father is having sex with with an animal and and and a child, an infant is there just watching the the the the scene. It is something that we never imagined that we would get contact with

• 14:24Speaker 01

  or get in touch with. Some of these things we tried as much as possible in as much as some lady very clear that we will be undergoing counseling. Counseling was not forthcoming, even after engaging ourselves in this text. Something that is worth noting is the fact that. From my position as the team lead, I was, for me, I, I, I am very good with interpersonal skills. I, I can tell when my team is not doing well. I can tell when they're, uh, when they're not interested in, you know, in, um, reporting to work.

• 15:08Speaker 01

  I can tell some of these things and this, uh, these were very, uh, clear indications that, you know, my team was just sending signals that they're not ready to engage with this, uh, such warnings.

• 15:23Alex Kantrowitz?

  Was this all of the text that you were seeing or like one out of every 10 pieces of text you had to read? How often would you see something, you know, as, as horrifying as the stuff that you just described?

• 15:35Speaker 01

  Out of, out of 10, you, a, a large percentage, uh, say about, uh, seven or eight, um, were more or less, uh, the same as what I had just described.

• 15:48Alex Kantrowitz?

  Wow.

• 15:48Speaker 01

  Um, three will, you know, will just be as neutral, but a large percentage will, um, will describe what I, what I indicated earlier.

• 16:02Alex Kantrowitz?

  Basically, it sounds like they were trying to train this model, um, to leave out some of like the most horrifying stuff. So they just kind of sent you guys anything that would be a, along the lines. And you helped basically determine what the line was that this trap, but, or this, this text generator would or would not cross.

• 16:21Speaker 01

  Absolutely.

• 16:21Alex Kantrowitz?

  Sound right to you?

• 16:22Speaker 01

  That's, that's the point.

• 16:24Alex Kantrowitz?

  Yeah. Did you see the text getting, you know, uh, better as time went on? Like, did you see it become, for instance, more coherent, more cleaner, like more clean? Did you think that the work that you were doing, uh, was, was having an impact?

• 16:40Speaker 01

  For me, the, the, the, the texts were not becoming any better, were not becoming any, any, um, uh, uh, holier, uh, so to say. Um, they were just becoming worse and worse. They, they, they were maintaining, at some point they maintained, uh, but they didn't become any better. So to a point where, you know, we say a sigh of relief, you understand? It was not, it was not something that you would want to, um.

• 17:16Speaker 01

  To engage with, you know, and

• 17:18Alex Kantrowitz?

  Definitely.

• 17:19Speaker 01

  For, for, for us and, and the team, we are actually at, at a point where, uh, between hard place and the rock, remember, um, we don't have any, uh, issue this, uh, uh, contract terminated. We, uh, don't know where to start from, uh, and we don't know, uh, we don't either want, uh, either do we want to, um, uh, proceed with this, uh, with this engagement.

• 17:50Alex Kantrowitz?

  Right. I mean, you had, had experience trading artificial intelligence. You knew what you were doing. So, you know, when you saw all this terrible stuff coming through, it's very different than labeling images. Were you just like, what, what the heck are we doing? Like, what exactly are we training? Like, did that thought come to mind?

• 18:06Speaker 01

  No, no, no, not at all, not at all. Remember, uh, at some point, the, the, the impression that we, we received, um, when we underwent trading was, you know, we, we will get these text messages, but we didn't, we didn't know the severity of how, uh, extreme some of these, uh, texts were. And again, you know, uh, for us, it was, it was a euphoria, just as I say, because we

• 18:40Alex Kantrowitz?

  You had work.

• 18:41Speaker 01

  We, we, we are looking forward to start work. It is during the COVID season, um, getting work, um, in, um, in a developing country is, you know, it's, it's a blessing in itself. You understand?

• 18:58Alex Kantrowitz?

  Yeah, I think what you're saying is that it was basically, it would even be a luxury to ask questions about the product. It was mostly like you needed, you needed work, this was work. And so you did it. And

• 19:10Speaker 01

  Absolutely.

• 19:11Alex Kantrowitz?

  Yeah.

• 19:21Alex Kantrowitz?

  You're the team lead. Do you go to SAMM management and say, "Wait a second, like, put us on a different project, or this is pretty awful stuff." What do the conversations look like after that?

• 19:34Speaker 01

  Well, for me, diplomacy really was key. So what, I could not even tell the management that we need a different project. My conversation was centered on, you know, we need canceling. The sooner the better. Because canceling is very pivotal in as far as what we are engaging with is concerned.

• 20:05Alex Kantrowitz?

  Right.

• 20:06Speaker 01

  At some point, the counselor reported one or two times, but you could tell he was not professional. He was not qualified, I'm sorry to say, asking basic questions like, "What is your name?" And, you know, "How do you find your work?" And for us, we need something like, "Please assist us," you know? "We need remedy right now. We need you to help us mitigate the attacks or the horrific situation

• 20:50Speaker 01

  that we have just encountered, you know, all through the week." You understand? But that was not forthcoming. So my job was centered on, you know, engaging the management to put mechanisms in place where psychological and social support will be forthcoming.

• 21:16Speaker 01

  But that be it as it may, it was deemed as if, you know, I was becoming, you know, authoritarian or not submissive. And it was not, it was not a friendly fire for them.

• 21:37Alex Kantrowitz?

  Right. Okay, let's go to break and come back right after this. We're here with Richard Matenga. He is a former team lead at Sama. So hearing that, you're gonna want to know why he ended up leaving and what happened after he brought some of these concerns to management. And then also, what did it feel like to see some of the product of this work after all this awful stuff, ChatGPT and OpenAI? ChatGPT becomes the most promising consumer app in a very long time and won't spend this stuff out. And OpenAI now is at the top of the world. So we're going to cover that and more in the second half. Back right after this. This episode is brought to you by Apppoint. Everyone's racing to roll out AI right now.

• 22:22Alex Kantrowitz?

  Copilots, chatbots, agents doing real work. But here's the part nobody loves talking about. All that AI runs on your data, and most teams have no single way to see it, secure it, and prove it's under control. That's exactly what Apppoint does. For 25 years, they've been the trusted layer beneath the world's most demanding data, now extended across your entire AI estate. Your data, your cloud, and the agents acting on your behalf. It's how more than 28,000 organizations deploy AI with confidence. So innovation scales without scaling risk. Apppoint, the unifying trust layer for AI. Head to apppoint.com to

• 23:04Alex Kantrowitz?

  see how enterprises deploy AI with confidence. Learn more at avpt.co/bigtechnologypodcast. That's avpt.co/bigtechnologypodcast.

• 23:18Alex Kantrowitz?

  Violent threats against executives are growing at an alarming rate. When a company experiences backlash, executives are the first people blamed. When their address and other personal details are sitting online, this backlash can lead right to their front door. The team at Ironwell knows this better than anyone. They've protected some of the most targeted executives and individuals on the planet for almost two decades. As a wide-glove enterprise security service, they protect your people with continuous personal data removal, proactive prevention tools, and human-led emergency support. So when someone goes looking for your executives, Ironwell ensures they hit a dead end. So here's what to do. Go to ironwall.com/bigtechnology, fill in the quick form, and request your free risk assessment.

• 24:01Alex Kantrowitz?

  Their team will show you just how exposed your executives are and how to lock it down before a threat reaches their front door. That's ironwall.com/bigtechnology. Learn how to stop online threats before they become real-world attacks. The link's in the show notes.

• 24:23Alex Kantrowitz?

  And we're back on Big Technology Podcast with Richard Matenga calling in from Nairobi, Kenya. And we're talking a little bit about the human part of reinforcement learning with human feedback. A lot of the reinforcement learning and the feedback gets talked about, not so much the human side, but we're here doing it. Richard, appreciate you being here. So when you spoke with your superiors about needing the counseling, they deliver counseling that's sort of subpar. What happens after that?

• 24:55Speaker 01

  So right after that, just as I said, we are in a friction with the management. And you know, my superior or my boss

• 25:11Speaker 01

  went ahead and wrote a report about me.

• 25:18Speaker 01

  That I was, I was insubordinate, that I was not quick in taking instructions. This was very interesting and very ridiculous, so to say. There was a project, and just put that at the back of your mind. There was a project that was coming, it was a lighter project, by the name ZF. And, you know, my, when it was, when they wanted to get resources to work on that project, the entire team that worked with me was picked.

• 25:55Speaker 01

  And my, my superior went ahead to write a report about me that I was not quick at taking instructions, which was very, very interesting. I engaged them to inquire, you know, how can this be possible? How can I, how can Richard be deemed as someone who is not quick at taking instructions? Yet, the people who are taking instructions from him were picked for this new project. It was very interesting, because they never responded to, to my, my inquiry. They never responded to my questions, but be it as it may, it, it never bothered me.

• 26:40Speaker 01

  What was at stake was, you know, the psychological imbalance that my team was, was experiencing. And I wanted that to be addressed as fast as it can. So that by the time they are moving to this new project, they are able to be, they are, they are in a position to, to say, you know. And this is where we we were, this is where we we are, this is where we want to uh to to be uh by the time

• 27:22Speaker 01

  by the time we are starting of this new project.

• 27:26Alex Kantrowitz?

  Right. By the way, so how long did you end up working on training the large language models, the text based models?

• 27:34Speaker 01

  The the that's very interesting. Uh the the period that we we were engaged uh with chat G P T was uh four months. And these four months, uh the the client provided a a a contract for one year. The the the contract between uh the client and Summer was one year. For us, we had a contract between Summer and between Summer and ourselves. The the client knew very well that we will be engaged for one year.

• 28:11Speaker 01

  But Summer, because there was another uh there was another prevailing um um story that was covered um or that was covered about content moderation, that did not portray the image of Summer and Open A I very well. Summer found it uh fit to uh terminate the contract between uh between them and and the client. So that they can protect the image uh so that they don't fight two uh walls at the same time.

• 28:48Alex Kantrowitz?

  What did your team say when you moved uh when when they moved, were were selected to move on to the project, but you were not?

• 28:58Speaker 01

  They were very frustrated, especially considering I I I I was, you know, every now and then I I kept on fighting for for for them, whether it is uh medical uh medical covers. I stepped in to to fight for, to ensure that they were uh they were covered medically. For those deferred uh deferred salary that I mentioned earlier on, I stepped in uh to ensure, you know, to ensure that the salaries were actually paid on time uh and not delayed. Um for any is- uh any issue that that prevailed, um I stepped in to ensure that, you know,

• 29:42Speaker 01

  um the the the issues were covered and they don't repeat, we don't see a repeat of the same issue over and over again. Because that's that's um the the the the things that I'm talking about were perennially uh happening um from month to month, from uh day to day, from week to week. And you could see that there was there was a commitment in trying to frustrate the team and the agents uh to a point where they are pushed to the wall, you know. But for us, our our patience was expressed in a very, in a very mature way.

• 30:23Speaker 01

  You know, my team was very mature in trying in expressing their their their patience. Uh it was it was my responsibility, however, to uh to face um uh the management and tell tell them, look, whatever you're doing is wrong and it it should not be happening uh in a a day such as this. It should not be happening to people who are vulnerable. My team was uh vulnerable. They were not being paid uh the salary that, you know, my su- uh supervisors have been paid.

• 30:59Alex Kantrowitz?

  Yeah, how much were they being paid?

• 31:01Speaker 01

  My my my team?

• 31:03Alex Kantrowitz?

  Yeah.

• 31:04Speaker 01

  My team uh my team's salary ranged between eighteen thousand, which is one hundred and eighty dollars per per per per month, to two hundred and two hundred dollars.

• 31:15Alex Kantrowitz?

  Hundred eighty dollars a month?

• 31:17Speaker 01

  Yes.

• 31:17Alex Kantrowitz?

  But you, how many hours did they work?

• 31:20Speaker 01

  Um that's nine hours.

• 31:22Alex Kantrowitz?

  Nine hours a day?

• 31:24Speaker 01

  Nine hours a day.

• 31:26Alex Kantrowitz?

  Five days a week?

• 31:28Speaker 01

  Five days a week, sometimes we we will even report on on on Saturdays.

• 31:32Alex Kantrowitz?

  Uh four weeks a month?

• 31:34Speaker 01

  Yes.

• 31:35Alex Kantrowitz?

  That comes out to a hundred eighty hours. So, are you saying that the salary for the folks who are doing this was a dollar?

• 31:44Alex Kantrowitz?

  Dollar an hour?

• 31:46Speaker 01

  Exactly.

• 31:47Alex Kantrowitz?

  Just give a sense as to like, is that, is the purchasing power in Nairobi for a dollar?

• 31:55Alex Kantrowitz?

  Let's say you wanted to buy like a, let's, let's say you wanted to go out for a meal at a restaurant. For instance, what does that cost?

• 32:02Speaker 01

  A meal will cost, averagely, it will cost um five dollars.

• 32:07Alex Kantrowitz?

  Okay, so it's, there is more purchasing power than a dollar in the US, but not dramatically more. I mean, not dramatically more, so it's equivalent to US dollars is, I mean, basically like paying people three dollars an hour.

• 32:24Speaker 01

  Yes.

• 32:24Alex Kantrowitz?

  In the US, which is, I mean, to me is, anyway, it's unacceptable. Okay, so what happens to you at, at Sama? Are, are you left, right?

• 32:34Speaker 01

  Yes, uh, so I left.

• 32:36Alex Kantrowitz?

  Yeah, talk about the.

• 32:37Speaker 01

  I have a friend, so I have a friend who, who is a director at Sama. He's the director of quality. Uh, I have a, I have the privilege of, I have had the privilege of growing up with him. Um, so, you know, we, we have that, you know, rapport, we have that uh, good and cordial relationship with him. However, uh, I came to know about, uh, I came to know his true self right after engaging him at Sama. You know, I will, I will, you will think that someone like him will fight for, fight for me.

• 33:17Speaker 01

  Uh, and, you know, and try and address issues related to agents uh very strongly. But that was not the case. So, you know, remember earlier on, um, before, before all this um, uh, happened, there was a prevailing case between, uh, sorry, there was a prevailing um, story that was being um, uh, told um, by one of the journalists um, uh, in the US, uh, in a lead, from a leading uh, media house. About the situation at Sama and Facebook.

• 34:01Alex Kantrowitz?

  Yeah, this was that time.

• 34:03Speaker 01

  Yes, so thi- this this was that time.

• 34:06Alex Kantrowitz?

  Hmm.

• 34:07Speaker 01

  So um my our our our story just came after. Uh came while this story was actually being been told or been drafted. So the the the director looked for me um with the insinuation that he didn't want me. Um to talk to uh to talk to any uh any journalists, because the journalist as well was looking for me. The journalist who covered the story about uh Facebook and and Sama uh was looking for me.

• 34:46Speaker 01

  So the director uh knew that um the the the journalist was looking for me. So he reached out to me, the director reached out to me um to find out whether I I got something. Uh or I got uh any engagement.

• 35:06Alex Kantrowitz?

  Hmm.

• 35:06Speaker 01

  But all through I have been lying to him, uh playing around with his integrity, telling him that, you know, yes, I got something. Uh you don't have to worry. Then he later on uh told me that.

• 35:21Speaker 01

  There is a journalist that is looking for guys who worked for your for your project, that is Chat Chat G P T.

• 35:31Speaker 01

  Um so I I I I asked him whether the journalist is the same person I know I know. Uh then he said, yes, he's the one. Then he said, uh I can see you actually on on the bench. Uh but if something comes up, I I you will be the first person we will we will consider. Then I said, um at the back of my mind, I I I am saying, this is a lie, because I've already cleared with Sama. I've already done my termination, uh my exit, uh yet I I am appearing on the bench.

• 36:09Speaker 01

  Um it is ironical, it is, you know, it is something which um is funny and hilarious.

• 36:17Alex Kantrowitz?

  So you were basically out at that point.

• 36:20Speaker 01

  Yes.

• 36:21Alex Kantrowitz?

  Mm hmm. And they were trying to put you back on the bench to like mess with your severance agreement? Or what was the idea there? Or just intimidate you?

• 36:30Speaker 01

  Yes, so um they they they they wanted to place me on the bench um so that I can you know, I can relax or or save the you know, I can be that person where every anything that comes my way, I it's a yes sir or a yes mom uh kind of response, you know?

• 36:56Alex Kantrowitz?

  If you're on the bench, yeah.

• 36:57Speaker 01

  If I'm on the bench, yes. So that I I will be I was I will be submissive and you know, and you take anything in, but that that that is not me.

• 37:10Alex Kantrowitz?

  Okay, so at the end of the day, you you decided to leave on your own.

• 37:14Speaker 01

  Absolutely.

• 37:16Alex Kantrowitz?

  And when was that? When were you actually full like officially out at Sama?

• 37:22Speaker 01

  So that was um April twenty twenty one.

• 37:28Alex Kantrowitz?

  Okay.

• 37:28Speaker 01

  Sorry, April twenty twenty two.

• 37:30Alex Kantrowitz?

  Twenty two. Okay, and about six months later, in November twenty twenty, Chat G P T comes out and next thing you know, it's the fastest growing consumer product in history. A hundred million people are using it within two months. What did you think when you saw that?

• 37:48Speaker 01

  To um a lot of things are are playing around in in my mind. You know, we need mechanisms put in place. Uh mechanisms that will protect content moderators um all around the world, whether it's chat G P T or or anything, you know? Um mechanisms that will help uh um the the the resource that is that are working on on on them uh become better and not uh not worse than how

• 38:26Speaker 01

  than how they studied it, you know, than how they were introduced to to to it in the first place. We need

• 38:35Speaker 01

  Problem solvers. We need people to identify with with um content moderators um just like any other professional. You know, these are professionals. These are uh individuals who are working. They are learned. Uh they are working uh on uh on uh such machines. They are placing uh information on machines and at the end of the day, they are chipping in enough uh hours um to their work. You understand.

• 39:08Alex Kantrowitz?

  Yeah. Was there any sense of like, oh, my God, like, have you tried ChatGPT? Was there a sense of, oh, my God, that product that we worked on? You know, does, you know, I I guess, yeah, I mean, was there any sense of like, wow, this was the product we were working on and and a reaction to see how many people uh actually wanted to use it?

• 39:29Speaker 01

  For me, uh and and for us, we we are very proud to to, you know, to have worked and uh uh for ChatGPT. We we we are very proud because we, first of all, what we did uh in as far as Summer is concerned, our project, uh Summer has this arrangement where it ranks projects as the best performing project. And for for for me and the team, we have ranked, um we have been ranking at Summer well for the period that we were there uh as the the number one project that, you know, that

• 40:10Speaker 01

  uh was ranked uh as the best performing uh in terms of meeting the KPIs. Uh so for for for me, even even in as much as um we will not, we will not uh tell how uh the project uh um went after after the contract was terminated, uh I can be proud of the fact that we we were the best, you know, and we did our best. Our quality was uh outstanding, our uh productivity was outstanding, our, you know, our KPIs, our attendance was, you know, above board. Uh, yeah.

• 40:52Alex Kantrowitz?

  Have you used ChatGPT yourself? And if so, what have you thought about the experience?

• 40:58Speaker 01

  Well, I have used it again. Um, um, the, the content that I, I went through is not as severe and it's not as uh, uh obscene as the one that we, we did at summer.

• 41:13Alex Kantrowitz?

  Hmm.

• 41:13Speaker 01

  Yeah, so the, the experience is completely different from, from the one that we did at summer.

• 41:19Alex Kantrowitz?

  Do you have a sense like, oh, like the work that we did, it worked, it was productive?

• 41:25Speaker 01

  Yes, we did, yes, yes. We, my, my team was the best performing team. It was the best that, you know, I could have ever, um, uh, asked or, or worked with. They were, um, they were productive, they were, uh, hardworking, they did that with, you know, uh, all the diligence that, that they had, you know, and they tried their level best. They did. The only thing, um, that, uh, the only challenge that they faced was, you know, little hand to, uh, hurdles, uh, here and there about deferred payments, about, uh, being rendered some form of counseling.

• 42:06Speaker 01

  Uh, this was disturbing, uh, to me.

• 42:11Alex Kantrowitz?

  Right.

• 42:12Speaker 01

  Yes.

• 42:13Alex Kantrowitz?

  So, so let me ask you this. OpenAI has raised eleven billion dollars and, okay, ten of that came after you had left, but eleven billion, one billion before. Why do you think they're coming to Africa and paying people a dollar an hour to do this work?

• 42:28Speaker 01

  It's very simple and your guess is as good as mine. The, the, the thing about Africa is that you have, there is a ready, uh, a ready resource, you know? There is a, a growing populace that is coming from, um, um, the campus and is coming from the universities who are saying, we need something to, uh, to engage ourselves with. The society in Africa, if you, if, if the population is not engaged, then there is higher level of, of crime, you know? Uh, people start killing each other, people start stealing from each other, people start

• 43:13Speaker 01

  doing all sorts of, um...

• 43:17Speaker 01

  Crimes, you know, and so the, the, the, the need to engage such, um, such pains, the need to engage such uh, resource is, is, you know, it's inevitable. So that's the reason why they, they, they, there is a growing supply in as far as uh, such uh, obligations or such uh, work is concerned.

• 43:45Alex Kantrowitz?

  Um, yeah.

• 43:48Alex Kantrowitz?

  A last question for you. How, how is your, your team doing and how are you doing in, in the months after uh, going through all this explicit material? And uh, are you, do you feel like you've gotten the support that you needed after going through what you have?

• 44:08Speaker 01

  Not, not, not at all, not at all. Um.

• 44:13Speaker 01

  The, the support that we, um, that we were looking for. Um, remember we have not even uh, received counseling ever since then. No one at, no one from Sama has, uh, has, has had the audacity uh, of reaching out to us. Um, trying to understand um, how things are or uh, how things are, uh, are going ever since we, uh, left. Uh, I can tell you for sure, uh, one of our colleagues, uh, um, uh, is, um, has had issues with his relationship.

• 44:56Speaker 01

  You know, he, he had issues with, uh, with, um, his wife and, you know, his wife left. Um, just because, um, of.

• 45:09Speaker 01

  The impact that that uh um the text that he was reading um had on him, you know, um it has never, it was not easy at all. It was not something that, you know, you will go home and say, I I am healed. I am better than how I started.

• 45:32Alex Kantrowitz?

  Yep.

• 45:33Speaker 01

  Yeah. And so I will, I will, I will try as much as possible um to um get to and know if all these individuals or all my team members are are are have gotten counselling and are getting counselling, that will be my, my, my, my joy and my happiness.

• 45:57Alex Kantrowitz?

  Richard, thank you so much for sharing your story. Appreciate you being here and I wish you well.

• 46:03Speaker 01

  Thank you very much, Alex, and looking forward.

• 46:07Alex Kantrowitz?

  And that'll do it for us here on Big Technology Podcast. Thank you so much, Richard, for coming on and sharing your story. I want to note that I've reached out both to Sama and OpenAI and have not heard back yet. So if they do respond, I'll include it in the newsletter that I'm planning for Friday. So hopefully we'll have a chance to hear from them there. You can subscribe at bigtechnology.com. And and yeah, hopefully we'll have some updates there. So more reporting is on the way. Um. So I hope you check out the newsletter. I'm so glad that you were able to listen to the podcast. So if this is your first time here, please hit subscribe. We do these twice a week flagship interviews on Wednesday and then we break down the news on Friday with Rajan Roy of Margins.

• 46:47Alex Kantrowitz?

  Um. If you are a long time subscriber and are willing, five star rating on Apple Podcasts or Spotify would go a long way. Thank you, Nick Watney, for handling the audio. Thank you, LinkedIn, for having me as part of your podcast network. Thanks to all of you, the listeners. Really appreciate you coming back week after week, sometimes twice a week. And I can't thank you enough. All right. Thank you, Richard, again for sharing your story. Uh. This is what the podcast is for. This is what we're here for, right? Bringing stories you're not going to read in the headlines, and I'm hoping that that's what we've got today. All right. That'll do it for us here today. Thanks again for listening and we'll see you next time on Big Technology Podcast.

• 47:29Speaker 02

  Now, more than ever, we need someone on our side. But Rob Whitman? Whitman's votes drove up costs, cut Medicaid, and risked 350,000 Virginians' health care to pay for tax breaks for himself and his donors. He's not on our side. Shannon Taylor will fight corruption, cut taxes, and bring down costs. This November, vote for Shannon Taylor. She's for us. Paid for by HMP, the House Majority PAC dot com. Not authorized by any candidate or candidate's committee.

• 47:58Speaker 03

  In this life, you can't win.

• 48:00Speaker 04

  When the world gets you down.

• 48:01Speaker 03

  No!

• 48:02Speaker 04

  Escape to Pluto TV.

• 48:03Speaker 03

  This is objectively fantastic.

• 48:05Speaker 04

  Relax into the movies and shows you already know and love, all for free.

• 48:09Speaker 03

  Shwing!

• 48:10Speaker 04

  Even our news is uplifting.

• 48:11Alex Kantrowitz?

  You stay classy, San Diego.

• 48:13Speaker 04

  Watch School of Rock, Crazy Ex Girlfriend, Anchorman: The Legend of Ron Burgundy, The X-Files, Wayne's World, and so much more.

• 48:19Speaker 03

  We're not worthy!

• 48:20Speaker 04

  The best things on TV really are free. You just have to know where to look. Escape to Pluto TV. Stream now, pay never.