On the search for action, Prime Video brings excitement. What did you think of shurken who talk instead of fighting in the last season of The Boys? You know, I can just pull you apart, limb by limb. And fighting where the birds never get off in the hilarious action thriller The Wrecking Crew. Yeah, I'm pumped. More epic action, on Prime Video you watch everything. The companies that are reporting high ROI and not the same ones who are reporting workplace reductions. So they're basically saying that those who are finding the value in AI are not the ones who are cutting their staff. In fact, in their belief, it's kind of the opposite. It does feel very much like a moment that could end up in unionization and could end up certainly in a lot more resistance from people within the tech sphere more widely, but certainly within Meta. Today on The Tech Report, I'm joined by the tech journalist, Natasha Bernal. Thanks for coming on. Thank you for having me. Algorithmic management, or in other words, using AI to directly monitor and in many ways control your workers, is kind of the next evolution of bossware people may have become familiar with during the COVID lockdowns. And now Meta is facing protests for its employees after it's basically told everybody to use mouse tracking software, which tracks clicks. But they also have done loads of keystroke software and things like that. These protests come just before Meta is due to lay off about 10% of its employees. Do you think that this monitoring software was the final straw that kind of pushed employees towards potentially unionization? I think it's difficult to say because they haven't said anything directly as sort of a result of this. The protests seem to be in the form of flyers that they've kind of stuck around different office infrastructure at Meta. But yeah, I mean, obviously these two things are probably linked. The idea that there's going to be a big amount of layoffs plus the idea of surveillance is not probably helping morale or the way people feel about working for the company at all. I would say that it's an interesting situation at Meta in particular because it feels like the workforce is at the moment split in two. You've got people who are being forced or encouraged to use AI as part of their job to make themselves more productive allegedly um and you've got the other side of people who are being surveilled by by ai um or kind of working towards certain goals that are set by ai or risk layoffs and so and this kind of feels like it consolidates both of those trends um where you're where basically meta is saying to workers we need to understand a lot better what people need in the workforce and how they can use our ai products better and therefore we're going to track what you do and sometimes take screenshots and often look at your keyboard and the way that you actively work so that we can then use that to train our AI and make it better. And I think that this is kind of the bridge that crosses the two, right? It does feel very much like a moment that could end up in unionisation and could end up certainly in a lot more resistance from people within the tech sphere more widely, but certainly within meta. To me, this kind of response seems somewhat inevitable given the layoffs and the recent treatment of tech workers. And certainly I think Meta has been a poster boy for bad AI, I'm not sure I'd say adoption, but forcing AI on your employees. How significant is it that Meta employees are protesting against their own company? I mean, it's pretty significant because of the history, as you say, like, you know, Meta has kind of a long track record of trying to get its employees to almost be the advocates and the evangelists of using its AI. And, you know, we've seen it go from saying, you know, we'd really love you to use our products to saying we'd like you to use any products. We would like to just, you know, watch you work and make sure that, you know, all of our tech is working. And it is a really interesting evolution to then say, now we want you to not just use it in your work and make yourself more productive, but we also want to watch the way that you work in general so that we can train our AI and you know I think in my mind's eye I could see that play out you know how the the mindset of maybe the corporate executives at Meta who thought this was a good idea kind of panned out in a room where they go you know we understand you know they have a wealth of data from Facebook from Instagram and even from WhatsApp on how your average consumer might want to use AI on how people might think can interact with their products. And that might lead to, you know, having solutions through AI that makes them better. But they don't necessarily have that wealth of data in the workplace. And so if they're trying to attract all of these corporate clients, which Meta very much is trying to position itself in that space, it needs to be able to gather that info. And I can imagine in the minds of these people saying like, oh, well, you know, we've got loads of people that work for Meta, why not use them. And I think it's just very interesting that clearly there's a clash of culture and of understanding going on here where, you know, people have clearly been like, look, we're okay with, you know, helping you train AI, we're okay with helping you create products and use products for meta, but we don't want to become the subject of that themselves. There's a line that's being crossed here that clearly some inside meta are very uncomfortable with. Last week, we talked a bit about google deep mind workers which were trying to unionize and how some groups from other other big tech companies had reached out for nothing specifically was said but in short kind of advice do you expect that we'll start seeing a wave of unions and maybe labor's rights movements and things like that especially given the the the large amount of ai is coming for your job and that kind of rhetoric that's happening at the moment yeah i think it's it is likely we're going to see more worker uprising certainly more worker resistance i'm not sure whether it will be in the form of unions i feel like you know it's still the vast majority of of people who work for these big tech companies are based in the united states and we've seen how big tech operates against unionizations and how aggressive that can be um i'm not really sure whether they'll form unions necessarily but it certainly feels like this uh discontent could lead to talent talent exit Certainly it could lead to people just deciding that they don't want to continue working there anymore, that they want to go somewhere where, you know, this kind of surveillance isn't happening and where companies align a bit more closely with people's individual values. And it does feel certainly like, you know, if you're looking to win a war, which is what these companies are trying to do, ultimately, they're trying to win this AI war. alienating your troops is one of the worst things that you can do at this point when you need that competitive edge I think the worst thing that can come from all of this this big movement of people saying we not happy with what is going on is the element of quiet quitting the element of them just saying actually we not going to quit We not going to leave We just going to do the absolute bare minimum possible and carry on, but we're not going to help you in your quest of winning the AI war. We're just going to sit and watch it play out and cash in our checks and wait until you lay us off because inevitably there'll be more layoffs. And I feel that that sort of very cynical view is something that we're likely to see a lot more of in the future. I want to carry on talking about the effect of bossware on a workforce and a workplace and that culture. But before we do, I briefly want to talk about a study from Gartner of about 350 global $1 billion businesses, which found that those executives of those businesses who cut staff after adopting AI or some form of optimization, automation sorry they saw zero financial gain compared to those who kept their employees during the AI adoption or automation process what what does that mean for for the promise of of AI productivity that's going to replace everybody I love this study because it's the first and what I expect to be a series of studies on what actually is happening behind all of these AI related layoffs the picture that executives have been painting for a long time now a couple of years has been one of we're making investments in AI and therefore we don't need insert x percentage of employees here right and that's been the narrative that they've been trying to push a lot of people within the industry and outside of the industry have been saying you know this is this is performative what you're trying to do is cut your costs on the bottom line and invest in AI and you can't do both you can't have people and have AI and so therefore you're trying to manipulate I don't know, massage the numbers. So it looks very much like you're making some cost savings, which AI is not currently able to generate. And this study kind of is the first one that proves that that is the case. So obviously, you know, Gartner has worked very closely with with people within the industry, just trying to find out and suss out whether, you know, these these big investments in AI have actually paid out, number one. And number two, whether these layoffs have actually resulted in like any positive, you know, ROI for AI that has been invested. And what they're seeing is that the companies, the company leaders that have been saying, you know, we'd like we'd like to make these cuts because we're investing in AI and are actually, you know, working on a very short term strategy. They're not going to necessarily see the returns that would replace all of the work that the people they've laid off were doing. and secondly they're kind of seeing that the companies that are reporting high ROI and not the same ones who were reporting workplace reductions so they're basically saying that those who are finding the value in AI are not the ones who are cutting their staff in fact that in their belief it's kind of the opposite so that the work the workforce reduction cuts were nearly equal for those reporting higher ROIs and those were smaller returns or even worse than outcomes on autonomous operations. So basically, they're saying these layoffs have not impacted whether you have a good return on AI or not. If you have a high return on AI, you don't necessarily have the equal amount of workplace disruption. And, you know, we've had a lot of people walk back a lot of the big statements of AI is going to take your job and say instead, it's going to turbocharge jobs. And this report very much cites that effect of saying, you know, we might actually see, at least in the shorter term, the creation of new jobs. And it's perhaps, I suppose, exposing in a very elegant way, that it's quite short sighted for companies to be claiming that they need to lay off all these people, when in fact, there might be a lot more work to do. We covered last week, basically, the effect of the loss of institutional knowledge and the experience that is lost by firing these people which is a whole nother conversation aside from the fact that there's just not the return on investment if you're firing people while adopting AI so go listen to that or watch that if you're watching and you're interested but moving on that Jevons paradox is something that comes up a lot on this topic is to explain why work or employee why there might be more jobs as a result of AI the the concept briefly is that rather than shrinking with efficiency gains workforce demand, workforce demands grow because there's a large appetite for whatever is being applied to. Op zoek naar actie, Prime Video brengt de spanning. What dacht je van schurken die praten in plaats van vechten in de laatste seizoensaflevering van The Boys? You know I can just pull you apart limb by limb. And fireballs where the bullets never hit in the hilarious action thriller The Wrecking Crew. Damn punked. More epic action. On Prime Video you'll see everything. A abonnement is scheduled. The audience can be advertised. 18+. The general requirements are of the use. Do you think this is what's happening here? Because to my understanding of this paradox there is... This effect is more of a medium to long term phenomenon. Yeah, I mean it's interesting because this theory is quite an old one. It's a 19th century one that kind of links to the Industrial Revolution. And the idea was that, you know, the advent of coal did not mean that we'd need fewer people in the workforce. In fact, we demand more people in the workforce. And I feel like here, the difference between AI and coal, there's many of them. But the big difference is here is the level of ROI. Right. And I think that the problem we're seeing at the moment is that the ROI is not necessarily enough to justify that paradox. and saying, okay, the Jevons paradox 100% works because we can see an increase in the amount of jobs. We're not seeing that yet. So I do think, you know, it is something that we could work towards that companies, especially as more of them start to see return on investments, that they might see that happening and therefore generate more jobs. It feels like it's a bit premature to say that that's happening now. But, you know, it is something that is very much being cited, which I feel is a much more positive turn to like the AI conversation than what we've seen thus far. It's all been doom and gloom for the last few years saying, you know, this is going to have a huge impact on us as a workforce, on us and our futures in terms of skills and on the way that we have to work. And so I think, you know, having that other conversation saying, actually, there's going to be loads more jobs is a positive one. And it's one that I feel more people in the sector should explore and say, what are the skills that might be good ones for people to start preparing for when that paradox is in play? So if we move back to the to the bossware stuff and what's going on with Meta, I'm sure lots of people watching will know very well what it's like being micromanaged and even having sort of every action monitored, maybe not necessarily by AI or some software, but at least by a helicopter boss. What are the benefits maybe of tools like these, of this BOST software? And do they balance the inevitable performative productivity that these kind of monitoring softwares create and has happened with Meta with their token maxing that kind of became more of a thing than I think maybe any of us would have wanted it to be. Yes, I think the alleged benefits to this new wave of AI tools is that they're supposedly able to detect and flag behaviours that might help managers to have more candid conversations with their teams. So, for example, they might be able to detect the language that you use at work the speed at which you're completing tasks during the day or even if they if they're able to do this if they're enabled to do this rather they can look at your facial expressions while you work and then they can tell or at least spot signs that you're suffering from really damaging burnout that you have too high a workload or that you're unhappy at work just based on on these sorts of things and flag them to your manager so that they can have conversations with you that's that's the theory right uh technically those tools might might provide a starting point i don't necessarily think that they might um aid in the in the in sort of solving these issues um but you know it might help theoretically to navigate times of really extreme stress reallocate work sometimes um and avoid people from quitting and having that attrition that costs employers a lot of money. In reality, though, I've not seen a lot of evidence that it's being used this way. The majority of surveillance equipment at work is used to penalise people. So the idea is that you have the knowledge, you impart the knowledge to your workforce that they are going to be surveilled, that they are aware, that they are going to be tracked in certain ways. And there is a responsibility from the employer to say that because it's only fair for the employee to be aware that they're being surveilled so that you know they're aware of the rules but also you have that situation where there's a lack of trust right you're establishing that you're watching this person and that should they not comply with the rules that you've set out whether it be they deem that the AI deems that you're not using language that's appropriate when you're customer facing whether it deems that you're not performing as fast as your peers or as well as your peers, there are consequences and none of them seem to be positive at the moment. So the AI is not likely to turn around and say, oh my God, Isaac, I've noticed that, you know, you seem to just be doing a lot of work and I'm just really sorry about that. Like maybe we should reallocate some work. I haven't heard of a lot of instances or any, in fact, in which someone has had that type of positive experience. It's been more like, you know, you're on a bell curve with the rest of your peers And should the AI determine that you're not doing as good a job, it will show up in your performance review and your performance conversations with your manager. And that's a lot more of a negative spin on surveillance, but it's much more what is actually happening in the field. We know that a lot of workers or some workers using AI in their jobs are experiencing burnout from, among other things, the increased managerial expectations and the demoralizing AI verification sort of workload replacing creativity. But what is the psychological impact of being managed by this kind of tool and the work environment that it creates? It's interesting because I think the biggest problem is the lack of trust. So when you as an employer decide to put in especially high levels of surveillance, especially AI powered surveillance, you're essentially telling your workers that you do not trust that they're doing a good job. you're also telling managers that you do not trust that they can manage workers properly and that they need some help and so that already there's like a unspoken rhetoric around that which it makes it very very difficult for people to enjoy their work and make independent decisions it puts that pressure whether you like it or not on people looking at their numbers looking at their metrics and becoming concerned about them. So you often find a situation where people are trying to game the system, not because they go from sort of doing a job that they are excited about or maybe don't enjoy, you know, there's a spectrum, but either way they start doing a job to suddenly looking at the system and saying, how can I position myself as best as possible within the system? And suddenly the game stops being doing a job well and starts being, you know, how can I allow this AI to rank me as highly as possible so that my job's not at risk, so that I don't get in trouble, so that I'm not perceived as being slow or less good as my colleagues. And this is a very, very easy thing to predict happening, but it doesn't seem to be something that anyone has found a solution to yet. It feels that the more pervasive the surveillance, the worse that situation can get. So if you've got somebody who is not only being, you know, tracked with keyboard strokes, which has been around since before the pandemic, in certain sectors, but certainly became a big thing during the pandemic. That's why you had those videos of people that had like, you know, something that was pressing the buttons every few seconds, or like people who are trying to game a Microsoft Teams to look like they were online so that it wouldn't go inactive if you didn't touch your keyboard. That has ballooned into something that's significantly different where you have people who are returning to the workforce or even starting in the workforce in office, I'm talking about, not just people who have started, but people who have been asked to return to the office full time, who are now facing a lot more surveillance, even in the office, than there was ever before the pandemic. And it's a really dire situation that a lot of people feel very uncomfortable about, but it's become so widespread. And this is the AI-ification of that, basically. you've got a situation where you they try to automate the surveillance and employers think that that's a really good idea but often the data that ai will have is not as good as as you know you might think and it might insert errors and might have perceptions about you that are not correct and what is happening at the moment is that employees and workers are in a situation where unless they keep extremely diligent records of their own it's very hard to dispute something that's watching you all day every day and say I'm sure I didn't take 10 minute breaks I'm sure that I was typing that entire time I'm sure that you know I had I didn't do this or I did do that it's very hard to have that argument and you're in a situation where you already know you're on the back foot because someone's watching you it's it's literally physically as if someone were watching you over your shoulder and I know everyone here and you know I'm sure people who are watching have had that experience and how difficult it is to type and and think and manoeuvre and do all the things that you would normally do because you feel like someone's watching you and so psychologically speaking I think it's really really difficult and the more pervasive it becomes the worse I think it will be. How can workers advocate for themselves when their boss maybe not their top level boss but the person or the the thing in this case that they are answering to is a black box which people say is objective or or neutral when when also we know that that is Beleef de kus die bijna gebeurde in de romantische film Love Me, Love Me. What is it about you? En een onverwachte romance die alle regels herschrijft in de serie Off Campus. Voor meer brandende liefde, Prime Video. Hier kijk je alles. Abonnement vereist. Inuit can advertenties bevatten. 18+. Algemene voorwaarden zijn van toepassing. basically not the case and couldn't be the case based on any data set having some form of bias in it yeah it's interesting because i think we've seen this play out before especially in the ride hailing and courier um side of of the of the employment sector um a lot of these a lot of people have um you know read all the stories about you know algorithmic decision making and that does take place a lot in that sector um and you know i've reported on it many times before but it's the idea of like ai will make mistakes much like human beings will make mistakes and and if employers don't put in an appeal system um or put in a a way for employees to to sort of stand up for themselves and be heard it's very very difficult for for things to go right so in in ride hailing good examples were for example um the ai that was not capable of identifying um couriers correctly if they were black or asian for example um and they would just say well that's not you and so therefore we won't give you work for the day and therefore people were going wait a second that is me and how do i prove that and there's no way of actually um you know contacting anyone and the people they were contacting were not responsive and then you sort of go well okay well then I don't have any work if you if you sort of transpose that example into the wider work sphere you can very much see a situation where imagine for whatever reason the AI has not recorded you going to work one day and you get into trouble with your boss and you say but I was at work and they're like well it says here that you weren't but then you know there's certain scenarios where you're not even dealing with the boss who do you who do you talk to you know if the AI says no, what do you do? And that's the big problem, I think, here with relying on a lot of AI that perhaps is making the decisions rather than just informing decisions, because you have no recourse to debate that. And, you know, anyone who's called into an HR meeting about their performance knows that, you know, the company's already kind of evaluated that data and made some decisions or made some suppositions about your performance. And so this just makes it a lot more difficult, the more power you provide to an AI versus a human manager means that you have less recourse to debate it and, you know, argue your side. And I think, you know, that's the biggest issue that psychologically people face. But also, you know, when you have a situation when you have AI placed as an objective evaluator, and it's not possible for anything to be an objective evaluator at all times um it's it's a tricky situation indeed just finally do you think this is something that could risk or maybe is already hardening work inequality because in researching some some things for this i i found that people that work alongside ai or use ai as the way that i would put it in their job day to day using it to do research or or whatever you whatever it would be are the ones which are getting wage growth and then on the other side of things thinking about warehouse workers or hospitality or retail or things like this they're maybe having shifts assigned by ai based on the productivity metrics which are all going on inside ai and being fed up to the management and especially for a product that was sold as democratizing power and helping people you can have access to a lawyer and all these sorts of things but maybe it actually like you say it's it's not so much that ai is the issue is that ai is being used to help make the decisions but that it is making the decisions and then employers employees rather have no recourse to that essentially yeah i mean it's interesting isn't it because as you say we we've mostly been talking about office workers but this has spanned you know everything from transport logistics to manufacturing so it's across every sector even you know people in the trades um also in a situation where they're being evaluated and i think it really this democratization no system is perfect but this democratization um is kind of assuming that the ai will be more objective than a human being um and and you know in certain cases i'm sure that could that could be the case if you've had if everyone has had a foul manager you might prefer for an ai to be evaluating you and managing you than a terrible manager that's true however i feel it makes a difference and who can who can see into that black box and who can probe the data that the ai is looking at and gathering and that does make a massive difference because again you know i posited the example of you know let's say the ai decides you haven't gone to work today like who do you talk to imagine if your manager and your manager's manager and your manager's manager manager have no idea what what is going on and they have to work on the assumption that the AI is always correct then it doesn't really matter even if you do have a conversation with someone if they're not able to probe that data and actually use it and and say for themselves whether this data is correct or not and you enter into a situation where it's Kafkaesque right it's one of those horrible nightmare scenarios that you might expect from Black Mirror where it's like the you know the algorithm has decided you've done this or you haven't done that and you know the only people who could possibly save you or the CEO of your company and the C-suite who are never going to look at your data and so yes you have a situation where a lot of workers have been fighting back again we've seen it in ride hailing we've seen it in manufacturing we've seen it in a lot of places where it feels that there are consequences if you do not hit a certain metric of performance that is analysed and evaluated by AI and you've got people who are trying to just game the system and it feels very much that instead of a democratization of AI, we're looking at a performative culture rising where people are just trying to desperately gain the system, not because they're trying to not do their jobs or be lazy or anything like that, but to try to gain the system so that the system doesn't penalize them and they don't end up at the bottom of the barrel. And that to me is not so much sci-fi, it's something that could very much play out in some workplaces where people don't have those human stopgaps and that human element to you know curtail a bit of the surveillance and ultimately I think employers need to really evaluate whether the surveillance that they're actually doing on their on their staff is justified because bums on seats was a strategy at some point now it's suddenly I can see through your camera while you're answering customer service calls, what for? Either you trust your workers or you don't, right? And I think, you know, that's a question that employers should be asking themselves, certainly, before installing any kind of technology that surveils people. Well, Natasha Burnell, thanks for taking the time. Thank you for having me. If you enjoyed today's episode and you want to hear more of The Tech Report, please consider liking and subscribing. Also, you can get episodes of The Tech Report wherever you get your podcasts.
