Utilizing Tech - Season 7: AI Data Infrastructure Presented by Solidigm - 2x29: How AI Can Help Displaced Workers with Saiph Savage

Episode Date: July 20, 2021

Although we usually focus on the ways AI can displace workers, this technology can also create new jobs and help them. In this episode, Saiph Savage joins Chris Grundemann and Stephen Foskett to discuss the many ways AI can help displaced workers. One new type of job created by AI is in the area of model training, and this can help develop digital skills and improve the lives of workers. Digital labor platforms tend to be opaque, however, and we must audit them to understand the wages paid, exposure to negative content, and invisible labor workers do to continue to use these tools. Yet despite these shortcomings, many workers report positive experiences, in terms of life/work balance, opportunity, and flexibility. Researchers like Savage are monitoring these opportunities and developing tools to help workers and policymakers fairly judge the costs and benefits of participating. Ultimately, these jobs can become a stepping stone to digital careers and further opportunities. References Saiph Savage’s Super Turker paper  Flexible Work and Personal Digital Infrastructures  Turker Tales: Integrating Tangential Play into Crowd Work  “Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass” by Mary L. Gray and Siddharth Suri (book) Three Questions How long will it take for a conversational AI to pass the Turing test and fool an average person? Are there any jobs that will be completely eliminated by AI in the next five years? Can you think of any fields that have not yet been touched by AI? Guests and Hosts Saiph Savage, Assistant Professor at Northeastern University. Connect with Saiph on LinkedIn or on Twitter at @Saiphcita. Chris Grundemann, Gigaom Analyst and Managing Director at Grundemann Technology Solutions. Connect with Chris on ChrisGrundemann.com on Twitter at @ChrisGrundemann. Stephen Foskett, Publisher of Gestalt IT and Organizer of Tech Field Day. Find Stephen’s writing at GestaltIT.com and on Twitter at @SFoskett.       Date: 7/20/2021 Tags: @Saiphcita, @SFoskett, @ChrisGrundemann

Welcome to Utilizing AI, the podcast about enterprise applications for machine learning, deep learning, and other artificial intelligence topics. Each episode brings experts in enterprise infrastructure and artificial intelligence together to discuss applications of AI in today's world. Today, we're discussing the ways in which AI can help displaced workers. First, let's meet our guest Saif Savage. Hi, it's such an honor to be here. I'm Saif Savage. I'm currently an assistant professor at Northeastern University. I also co-direct the Civic Innovation Lab at the National Autonomous University of Mexico, UNAM. And within these two universities, what I'm doing is that I'm using AI to empower workers. And as often, I am Chris Grundemann, your co-host. I am a consultant, coach, mentor,

and content creator. And you can learn more about what I'm doing at chrisgrundemann.com. And as always, I'm Stephen Foskett, organizer of Tech Field Day and publisher of Gestalt IT. You can find me here every week on Utilizing AI, as well as on the Gestalt IT Rundown every Wednesday. So one of the things we've talked about quite a lot on Utilizing AI is the many ways in which AI can displace workers. In fact, that's one of the three questions that we often ask our guests is asking them about positions that may be completely eliminated by AI in the future. We also sometimes talk about some kind of dystopian stuff here about how AI is running roughshod over human rights and the ethical and moral boundaries of using AI.

That's why I was really excited to get a chance to talk to Saif about this, because frankly, you're bringing a little bit of the sunny side of the street here, because rather than talking about how AI can displace workers, you're talking about how AI can help displaced workers. Can you give us a little bit more background there? How can AI help workers who have been displaced? Yes. So the first thing that we need to understand is that AI has also been creating new types of jobs. What do these jobs look like? A lot of these jobs focus on getting human workers to do the tasks that are difficult for the machine to do. So for instance, you might have human workers who are going to be labeling images. By labeling these images,

these labeled images are then going to be fed into a machine learning model. And so the machine learning model will be able to understand, oh, this is a stop sign. This is not a stop sign. And through this process, a self-driving car will be able to better understand the world around them. Other types of tasks can involve, for instance, categorizing content by saying, oh, you know what? This Facebook post is filled with pedophilia. This Facebook post has a lot of hate speech.

Facebook's recommendation algorithm is not going to show you content that is criminal or that has disinformation or that might be extremely violent. And you'll be able to get content that is better focused on what you like. Other types of tasks that these workers can do can involve, for instance, transcribing audio. This way, Alexa will be able to better understand users who have different accents. And so AI is creating new types of jobs. And what my research has focused on is training, for instance, workers from rural communities who have been displaced to help them to develop their digital skills so that they can now take on these new types of jobs that AI is opening up for them. Awesome. That's really, really interesting. And I think it's,

as Stephen pointed out, right, it's kind of one of the leftover things that people don't talk about that often, which is that AI, at least it seems to me right now, in the current phase of AI adoption, it seems to be creating more jobs than it's destroying. It's definitely affecting a lot of jobs, but it's also creating a lot of jobs. However, I do want to take a moment without going too far off track here of the sunshine and rainbows part of the story, just to talk a little bit about some of those jobs, you know, maybe the quality of those jobs. And so Saif, I'd like you to talk to us a little bit maybe about, you know, I've heard a lot of, you know, this like shadow workforce, or, you know, we've heard the term mechanical Turk, and there's definitely a lot of human involvement in a lot of

this, right? And, and I guess, you know, for one thing, right, you know, are these jobs paying better than the alternative? Or is this exploitative in some cases, right? Just like we saw with manufacturing, offshoring and, you know, physical jobs went to other countries where maybe people weren't being treated well. Is that happening in the AI space? And then also, if you're doing like content categorization, right, is there a long-term effect of, you know, looking at terrible images and classifying them as such over time? Yeah, those are great questions. So one of the things to consider is that what is happening inside these digital labor platforms where now we have workers that are helping the AI, a big issue that exists is that we don't know what is going on. For instance, we don't know what type of wages do workers have, just exactly how much are they exposed to this type of violent content.

And so part of my research has been developing auditing tools through which we can start to understand, okay, what's happening inside these platforms? What type of wages exist? So one of the things, and so on one hand my research lab developed plugins through which we can measure what is the hourly wage that workers have, that workers on these platforms, particularly Amazon Mechanical Turk, what type of wages do they have? Another thing that we recently just developed is a web plugin to be able to measure the amount of invisible labor that workers have to do on these platforms. Invisible labor is work for which workers are not paid, but they have to do that work in order to survive on the platform. And so right now I'm developing a lot of auditing

tools to be able to understand what is happening. So on one hand, my research found that a great number of workers on these platforms were earning less than minimum wage. They were earning around $2 per hour. Now, what is also important to consider is that there are also a number of workers who have been able to thrive on these platforms. And so my research has focused on understanding how exact, what exactly are the workers who are thriving doing so that we can help the new workers develop those types of digital skills so that they can also thrive.

I think some important things to consider from these platforms is that on one hand, the platforms help workers to not have to abandon their hometowns. For instance, in rural West Virginia, we had suddenly a number of ghost towns that started emerging. Why? Because workers needed to go into the city to find jobs and abandon their hometowns. With these types of platforms, workers can now stay at home and continue growing within their community. Additionally, within rural communities, also instance, working on these platforms because it allowed them to be able to take breaks whenever they wanted to be able to take care of their

children and then make a living. And so, for instance, for rural communities, these types of setups were highly beneficial. We also found that these types of platforms were also beneficial for people who had different types of disabilities. Why were these platforms beneficial? Well, on one hand, workers could be operating from any type of setup that they wanted. Many times offices would not allow them, for instance, especially if they had chronic disease, to have that type of setup. Another thing is, for instance, people who had depression expressed that they found that type of work highly motivating because they could do small tasks and that helped them to feel accomplished. what problems are these platforms bringing and have a quantitative way of measuring those types of problems so that we can then help policymakers to be able to make these platforms accountable.

For instance, through my tools, we've been able to identify that, hey, Amazon Mechanical Turk to a number of workers, it's paying workers less than minimum wage. So I think that providing these types of tools is also important to provide change. In a way, it really reminds me of what we're hearing about the pandemic and the sort of work from anywhere trend, and as well as, of course, the sort of task worker trend that we've seen throughout the economy. In other words, whereas, you know, you could easily criticize, you know, not to name a particular company, but, you know, like a car sharing service, for example, for being exploitive of workers, you could also, frankly, find some workers who are very, very pleased with the

opportunities that it's provided them in terms of flexible work or something like that. basically get paid as much as they work, which is something they weren't able to do previously in regular, you know, regular jobs where, you know, you basically have to, you know, abide by the work hours and the, you know, the restrictive schedules that you're assigned. They really appreciated the fact that they could really kind of work as much as possible. And I actually have a good friend who is one of the workers you're describing, who is doing this kind of AI model training on a piecework basis. And this person is extremely excited about the opportunities that have been provided, even though, quite frankly, they're also quite disappointed with some of the, let's say, content that they have to work through in order to train the models. And so if I can understand,

Saif, I think what you're saying is that with proper supervision, with proper auto monitoring and metrics and reporting to authorities, these platforms can actually be quite empowering to disadvantaged people. Yes, correct. I think that we need to understand in more detail what are the benefits, what are the problems as well that these platforms have, and also respect, for instance, people who are enjoying the work on these platforms. For instance, a number of workers also expressed that they found the job, even though you could consider it monotonous, tedious, they particularly liked it because they could zone out and do work and make money. So for instance, we recently interviewed actually a web programmer who was in Russia. And so he was a constant worker on these platforms.

And he personally said that he liked the fact that he could kind of just, for instance, after a long day of programming, get on the platform, kind of zone out and do tasks that were not as complicated and make a living and just gain extra money. And so I think that it's also respecting, for instance, that people have different types of lifestyles and some lifestyles having that opportunity, as you mentioned, of making additional money, having a job that allows you maybe to be able to zone out because it's not as a complex tasks, for instance, as maybe programming. I think that that's also important to to identify but we definitely do need tools through which we can audit what is happening so that we can identify what type of changes, should we also aim for. I really like that approach of looking at the, you know, the incidences where this is working really well, right? It's the positive deviance approach, right?

As a side note, if nobody knows about Jerry Sternen and the Save the Children work he did in Vietnam, it's a really cool story. And it kind of led to this whole movement of positive deviance, which is basically, instead of trying to find the problem and solve the problem, you find areas where the problem doesn't exist and figure out why, and then spread that knowledge. And I think looking at folks who are really thriving on these platforms, and then finding ways to replicate that follows that pattern and really resonates with me. And so I wonder, you know, going, you know, further down that path, I mean, obviously, you've laid out some really good examples of kind of situations where this stuff makes sense. Um, is your research far enough along to start learning, you know, what those, you know, what those positive deviants have in common and, and what can make folks successful. And then how does that get applied back into, you know, you talked about training

and things. So what's the next steps here and how does this look? Yeah, actually, um, we, I recently have a two research papers around, one is called Becoming the Super Turk, which is about how we can design tools that basically learn what are the patterns of the workers who are succeeding, and then guides other workers to follow those strategies so that they can also thrive. Some of the patterns that we're identifying actually is that the expert workers are very good at identifying what tasks are not worth their time. That's one of the main differences. So the novice workers will take on tasks that are not paying them well,

but they haven't realized that the task is not paying them well. And so they'll do the task. And then while they're doing the task, they'll suddenly realize like, oh, this task I thought was going to just take me, let's say three minutes. It's taking me half an hour. And for the amount of time that I signed up for this, it's not worth it. One of the things to consider is that work on these digital labor platforms usually tells you how much they're going to pay you for a piece of work, but they don't tell you much exactly does a piece of work usually take so that we can inform workers about that so they can make better decisions about, okay, is this, and we then also predict approximately what's going to be the hourly wage that they're going to gain if they do those tasks. And so through this, we're helping novice workers to be

able to better navigate this space. Yeah, that actually reminds me, though, of another criticism that we've heard about a lot of these contract workers out there in the gig economy. And that's that, as you say, many of them aren't aware of the actual take-home pay that they'll be earning from this, both because it's difficult to understand sort of how much the per unit cost translates into a per day, week, month, year earning potential. Also though, there are costs associated with this? You mentioned the invisible work that happens, but of course, there's also costs in terms of equipment and services needed and so on. So I wonder, are you looking at that as well? So for example, do people need internet access? Do they need equipment that they might not have access to? Yeah, that's actually a great point. So with an invisible labor, some of the things

that we were looking at was, it was particularly for the platform of Amazon Mechanical Turk. So one of the things to consider is that these digital labor platforms have put onto the shoulders of workers costs that were traditionally absorbed by companies. So for instance, when you were in your office job and you were deciding what was going to be the next task that you were going to do, you were still getting paid. Even if you were at your desk just planning off your day, for that time you were usually paid. For workers on these platforms, for instance, finding what type of tasks they're going to do, this is time for which they're not getting paid. Similarly, in your office job, for instance, if you had to email

your boss or email maybe the secretary to comment out some things, that was, again, time for which you're getting paid. This is also time for which workers are not being paid. So a lot of those costs used to be absorbed by the company, for instance, Microsoft, to their engineers, if they're deciding what they're going to be doing next, if they have to send out some emails, all of that is time for which they're getting paid. Now, in these new digital labor platforms, those costs have gone on to the workers. And so, workers, for instance, do not get paid for messaging the employers that hire them. They do not get paid while they're searching for tasks. And so this is now a cost that workers have to absorb. And so that's right now what I was measuring. Within the platform of Amazon Mechanical Turk, what type of invisible labor exists and starting to quantify it. The book

Ghost Work, which I highly recommend, is from one of our co-authors in this new paper that we have where we're quantifying those amounts. And that book provides a really nice overview of the different types of invisible labor that exist on platforms such as Amazon Mechanical Turk. And with respect to the resources, one of the things that we're also considering is how we can leverage public infrastructure to empower workers. So for instance, part of my research has focused on helping rural adults be able to develop their digital skills so that they can now get on these platforms. A big problem that exists, as you know, is that, well, maybe they don't have access to internet. Maybe they don't have the computers. So here I'm teaming up with public libraries in rural areas so that we can transform the library into spaces where workers can come in and they can start to develop their digital skills. It almost seems like there also needs to be like business skills, right?

Not just the digital skills. And what I mean is a lot of the things you're talking about as we talk through this, right? This invisible work, things that you have to do to keep yourself afloat doing this piece work that you're not necessarily paid for. You know, to be honest, it sounds a lot like self employment, which is an area where now all of a sudden, there's all these additional tasks you have to do that are the responsibility of your business, which is you that before someone else took care of. And and I can definitely see just just some savviness around that, you know, even outside of the digital sphere. So is that a piece of it too, or is it just the

digital literacy, or is there also just kind of a, I mean, it's almost prioritization and time management, you know, included in here, right? And looking at this as if you're running a self-employed business, kind of. Yeah, completely. Actually, those types of skills have been coined gig literacy skills. So gig literacy skills is all about, for instance, how do you present yourself on your worker profile so that employers will hire you? How do you respond to certain employers

so that they'll be happy with you and maybe give you a bonus? What type of employers should you avoid because they're scams? So a lot of those are definitely skills that workers have to develop in order to be able to better navigate the space. One of the things that we've been arguing in my research is that we need personal management infrastructure that will help workers, for instance, to be able to navigate the different online spaces

where they are working, especially because as you mentioned, for instance, maybe within one digital platform, they've been really good at learning how to present themselves, how to manage their time. But suddenly if workers want to transition to another type of digital labor platform, how can we help them to bring in those skills that they developed in the previous platform

so that they can also succeed in the new one? What type of skills are transferable so that they can also thrive? I have a new paper as well on this topic about providing workers with personal infrastructure through which they can achieve their different goals. And that's actually, I think, one of the things that I hadn't really considered until you mentioned it earlier in this discussion as well, is the empowerment that comes to underprivileged people and people in rural communities, minorities, you know, people who have not had the opportunity to participate in the digital economy. Many cases I could see now that these jobs can open doors to empowerment, to digital empowerment for many of these people. So,

you know, perhaps they're starting off by doing, as you said, mechanical Turk tasks, but in order to do that, they have to develop skills related to computers, you know, they have to build up, you know, find connectivity, whether it's at their home or at a library, for example, but then, you know, build up that kind of digital literacy, computer literacy, and that can actually improve their lives in other ways. So perhaps should we be seeing these as stepping stone jobs that can help people to come, you know, first they're just doing, you know, digital piecework, where we're taking, getting these workers on these digital platforms as a stepping stone for then getting, for instance, office jobs. One of the things to consider is that it's, I think that a big challenge that we have within this space is understanding what are the requirements that the different employers have in order to accept workers. So for instance, you might have workers who develop really good skills on Amazon Mechanical Turk that help them, as you mentioned, learn where they can get internet access, become faster at typing. And so they could then, you could think, well, you know what, you have the skills to now be able to take on an office job in West Virginia. The problem that

we're seeing right now is that workers have different mental models about what those skills mean and what employers are also asking. And there are also differences between local employers and, for instance, employers that are more global. For instance, let's say you have a company like Amazon or Microsoft that could hire these workers, the requirements, how Microsoft, for instance, is presenting the skills that workers need is going to be different than how a local employer in West Virginia is going to request those skills. And also different for how, for instance, a community college is presenting the skills that they're helping workers to develop.

And so I think that part of it is providing tools through which we can facilitate those transitions and also inform workers about how do you brand yourself so that the skills that you developed, you can present them based on what the employer, for instance, Microsoft is looking for. I see some corollaries here, and I know we kind of made this, or maybe I made this jump earlier in the conversation, right, between, you know, offshoring of factory jobs and and, and the perceived exploitation or actual exploitation

that has gone on during that. And the corollary here, and what I mean is, you know, in the long term, while there have been definitely sweatshops, and many things have gone wrong with with offshoring and manufacturing in other countries, there's also been some things that have gone right, which is that some of these countries, China included, have built a, you know, massive and growing middle class from folks who were able have built a massive and growing middle class from folks who were able to make a living and start saving money working in these factory jobs who are now in this generation, we know a generation later, have actually taken some of that money and are opening factories of their own.

And I mean, just based on the conversation Stephen and Saif just had, it sounds like that's a potential here as well in the digital world where you kind of, you know, open up this opportunity for folks to get involved in this gig economy, which potentially has some compound interest to be paid to future generations. Yeah, I think that part of the problem, I really like this idea about how we can, for instance, help certain individuals be able to transition to maybe higher classes, etc. I think that one of the things that we can build on is looking into the past about how this has worked out and identify some of the strategies that we can incorporate. So for instance, here I've been reading a lot, a critical theory author called Marcuse, who during around the 1940s, 1960s, he was looking a lot into the working class and the fact that maybe our current world was pushing us into accepting a certain reality where, for instance, maybe you would have, let's say, the middle class or lower classes stuck in jobs where they are not going to be able to

evolve. And they get stuck in what he was calling a one-dimensional thinking where we only see the world under one lens and we only see that one type of reality is available to us. And so we could argue that the problem with these current platforms is that we're enslaving, let's say, these workers into believing that they can only stay on these platforms forever. This is the only reality that they're ever going to see. One thing that Marcuse says that we can do to start to fight, for instance, against that one-dimensional thinking is through fiction. So by integrating fiction, you can power to define new types of digital labor platforms that they would like to see, and also how do we help them to make those ideas into reality. And so on one hand, we developed a tool called Turkertales, and I also have a research paper

on that, where we studied how we could integrate play and fiction into the digital labor platforms that workers are on. And then I also developed a tool called MetaGig, which allows anyone to be able to define the digital platform that they want for themselves. And so through this, we're helping workers to become entrepreneurs. And as you mentioned, Chris, that can be actually pretty exciting because you could also consider that workers might be right now identifying important problems that exist with current platforms, problems that we don't even see, and opportunities about how these platforms could be different. And so now I'm providing them the tools to start to reimagine and build new futures? It really is. I have to say, I really hadn't thought of a lot of these things.

And it really is kind of eye-opening to think about these things in a more positive way. Because frankly, we spend a lot of time criticizing everything that comes out of the digital economy and the gig economy and artificial intelligence. And sometimes we don't stop and step back and look at the ways in which it can help people. And I have to say, I really hadn't considered many of these aspects and the empowerment that can come to people. You know, just off the top of my head too, I'm thinking, you know, how would a rural, you know, person in a rural part of America, not only would they maybe not be able to describe themselves to a high-tech company, but they wouldn't even have a line of communication to

that company. But yet some of this kind of digital gig work can automatically open up that line of communication that can present them for work that they might not have otherwise ever been able to find. It just really is remarkable to think about all the aspects of this. Chris, what do you think? Are you convinced? Do you think that there are some blue skies behind all these gray clouds? Absolutely. Yeah. I mean, I think we still need to be very, very careful about the current realities of some of this invisible work and some of the exploitation that might be going on or could happen. But I definitely, you know, leave this conversation much more optimistic about the doors

that these same platforms can be opening for folks. I mean, it's really exciting. And thinking about how to, you know, capture that skill, translate that skill into a true resume, open those doors. And whether it's to more advanced gig work or a career or whatever people want, because I think that's changing quite a bit right now. At least it seems to be. But making that sustainable in a way where whether they stay a gig worker or move into a career, they can feed their family in a way that maybe they couldn't have without this opportunity. It's amazing. Yeah. And similarly, as somebody who really enjoys history, I think as well, I hadn't really considered the corollary between digital gig work and the factories that sprang up across the United

States in the 20th century, which similarly, I mean, they were challenging, of course, and they were not a good experience for some people, but they did open some doors, especially to underprivileged people who might not have otherwise been able to work. And if you think about all the people who went to work during World War II, for example, or during the economic boom in the 1950s and 60s,

who would not have been accepted by society, but were accepted because they were needed on the factory floor and improved their lives that way. It is a blue sky in an area that obviously looks like gray clouds from the outside. So thank you so much, Saif, for giving me so much to think about and hopefully giving our audience so much to think about. So as we warned you at the beginning of the recording, now is the time when we transition into our three questions segment where we ask you three open ended questions that, frankly, you haven't been warned about, but hopefully it'll give you a chance to think on your feet and surprise us with the answers. So we have a set of these questions and I try to match them to our guests. This time, well, let's see where we go here. Let's start off with the obvious question. One of our three questions is asking if there are any jobs, people's jobs, that will be completely

eliminated by AI in the next five years. Now, you, in your position, maybe have a different perspective on this, but even so, are there jobs that are going away because of AI? So some of the jobs, for instance, relate a lot to even just you could think about some of the people in libraries, for instance, that were maybe recommending and finding you certain books. I think that that type of job right now is being transformed into new type, I think that the library, for instance, especially is transforming itself. Now, I could actually see the library as spaces where people can come in, and as I mentioned, develop their skills. And so maybe instead of having a library person who is recommending you

books and helping you find books, this might be a person who is maybe more helping you within your different career growth or helping you develop those types of digital skills. And so I see a lot of, I think also a lot of repetitive tasks. For instance, we used to have as well people who would be in the Q&A, for instance, from cities. Those are jobs that right now are being highly automated. Jobs that will likely not be as easily automated are jobs that involve a certain creativity. And so that's why, for instance, integrating fiction can be important because those types of jobs are not going to be easily displaced. And I think especially, for instance, if there are managers and startups, CEOs, basically, who are listening to the podcast, what I would advise them is to think about from the workforce that they have, how can they help their workforce to transition to new types of jobs? For instance, the person who would use to help you find certain books, this person

might actually be very good at providing maybe certain context about a certain recommendation. So you could think about this person now working alongside AI to provide contextual information about the recommendation that AI made or even help correct the AI when it's recommending certain books that don't make sense for certain topics or complementing AI where it has gaps. So I think identifying those gaps can be useful. Excellent. And my sister, the librarian, would absolutely agree that librarians are doing all sorts of things that you don't think they do already. So next, you know, you talked about communication. How long do you think it's going to be before we have an AI that can talk back and fool you into thinking you're talking to a real person? So I would say right now we do have this,

especially when we think about human-centered, sorry, crowd-powered AI. So for instance, one of my collaborators developed an agent called Chorus, which matched AI with crowd workers. And so when the AI failed, the crowd workers would enter the conversation and help correct the AI and also provide different ways for the conversation to go. So I think that these types of hybrid systems that are mixing AI plus these crowd workers right now would completely fool you into thinking that you're talking just with a human because it's taking the best of both worlds. Interesting. All right, let's wrap it up with this one.

Can you think of any fields of human endeavor that have not yet been touched by AI in any way? It always takes a moment. Wow. Maybe part of it is when we're thinking, so I know for instance that right now we do have AI that's learning a little pieces from the indigenous culture, which I haven't seen AI be involved, and precisely because it involves a lot of different human creativity that might be hard for AI to capture. But yeah, that's a tough question, actually. And also, I think here, maybe it's also thinking about where we don't want to see AI. And for instance, it's also valid to say that there might be jobs where we prefer to always allow humans to be able to take them on. For instance, maybe part of being maybe a psychologist for people, maybe some tasks can be delegated to machines,

but at the end, you do want to have a human component to provide that type of support. So I think it's also important to think about what type of jobs would we not want AI to be involved in? I should have known that you'd come up with something that I hadn't thought about. That's great. Thank you.

Think about where we don't want AI to touch. Well, I really appreciate you joining us today. I have to say, this has been a really an eye-opening episode and it's been wonderful to learn from you. Where can people connect with you and follow your thoughts on AI? And maybe is there something that you've done recently that you would like to direct them to? Yeah, please visit my website. It's saif.org. And here you'll find a lot of the different research papers that I discuss. For instance, check out our work on quantifying the invisible labor on these digital labor platforms. You can also find me on Twitter. I'm Saif Sita. And there I'm sharing a lot of different data about how we're empowering workers, also how we're working with different governments to also empower workers. You can also visit, yeah, those are my main sites. Thank you so much. It was such an honor to be here. Great. Thank you. And Chris, how about you? What are you working on these days?

Yeah, working on lots of stuff. There's new papers being published by GigaOM and also elsewhere. You can find everything to do with me on chrisgrunman.com, although definitely open to have a conversation on LinkedIn. And you can follow me at Chris Grunman on Twitter as well. And as for me, you can find me here at Utilizing AI every Tuesday and on the Gestalt IT Rundown every Wednesday. And I really appreciate the opportunity to have these conversations with you and with the rest of our community. So please do look me up on Twitter, connect with me. Thank you very much for listening to the Utilizing AI podcast. If you

enjoyed this discussion, please do give us a rating and review on iTunes, since that does help. And also, please do share this episode and the others with your friends and colleagues. This podcast is brought to you by gestaltit.com, your home for IT coverage from across the enterprise. For show notes and more episodes, go to utilizing-ai.com, or you can find us on Twitter at utilizing underscore AI. Thanks and we'll see you next week.