# AI has not produced broad white-collar job losses

Markdown copy of https://realalpha.app/s/ai-has-not-produced-broad-white-collar-job-losses/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

- Speaker: [Ed Zitron](https://realalpha.app/people/ed-zitron/) · Ed Zitron is a British tech critic, host of Better Offline, and founder and CEO of EZPR. · Hosts Better Offline and writes Where's Your Ed At, a prominent critic of AI industry hype.
- Host: [Steven Bartlett](https://realalpha.app/people/steven-bartlett/) · Host, The Diary Of A CEO
- Show: The Diary Of A CEO
- Episode: [The Man Who Calls BS On AI: They’re LYING About AI, 2027 Is When It All Breaks! | Ed Zitron](https://realalpha.app/v/Lf5oqGOCRCM/), Aug 27, 2026
- Moment: 1:00:38 to 1:07:33 of the episode (7 min), playing on https://realalpha.app/s/ai-has-not-produced-broad-white-collar-job-losses/; the episode from there: https://realalpha.app/v/Lf5oqGOCRCM/?t=3638

## Quote

> “work-wise, are art directors. It's people, art directors, transcribers, translators, who have bosses that **don't care about the output**.”
> — [Ed Zitron](https://realalpha.app/people/ed-zitron/) at [1:06:31](https://realalpha.app/v/Lf5oqGOCRCM/?t=3991) · The Diary Of A CEO

## Takeaways

1. Autonomous driving still faces difficult edge cases and slow deployment.
2. He says job losses are concentrated in low-cost creative and contract work.
3. He cites no link between AI token spend and revenue per employee.

## About this segment

Ed Zitron argues that autonomous driving and white-collar automation are being oversold, and says job losses are concentrated in cheap contract work. He cites a study finding no link between AI token spending and revenue per employee.

From “The Man Who Calls BS On AI: They’re LYING About AI, 2027 Is When It All Breaks! | Ed Zitron”, Aug 27, 2026. [Watch the full episode](https://realalpha.app/v/Lf5oqGOCRCM/)

## Transcript

From the transcript The Diary Of A CEO publishes with the episode.

[1:00:51](https://realalpha.app/v/Lf5oqGOCRCM/?t=3651) My car that drives itself, that is AI technology. Yes. I sat here with Dara from Uber, and he was saying that, I think in a couple of years' time, we won't need drivers for Uber because the cars will drive themselves. Like, they'll be fully autonomous. And I think, if I'm not mistaken... Driving is one of the biggest professions on planet Earth. So when you hear people, when you hear the CEO saying that there will be job disruption, you say that they are not telling the truth.

[1:01:22](https://realalpha.app/v/Lf5oqGOCRCM/?t=3682) Yes. Or they're guessing in a way that's very good for them. Think about it from the perspective of Microsoft. Satya Nadella. He's not going to be like, yeah, we're done after. This is going to work, mate. Of course he's going to talk his book and he's going to say, yeah, this is going to replace all workers. It's going to be amazing. It's going to be so powerful. And then he'll change his tune and say, actually, it's not going to replace workers.

[1:01:40](https://realalpha.app/v/Lf5oqGOCRCM/?t=3700) It'll make them more powerful because the things aren't catching up. They're from Uber, for example. Of course he's going to say, if this happens, then that would be good for Uber because Uber would just become... An autonomous taxi service. There's a reason that Waymo's taken... I find Waymo fascinating. I think that shit's really cool. I think there are socioeconomic problems that will come from it.

[1:01:59](https://realalpha.app/v/Lf5oqGOCRCM/?t=3719) I think there are actual real problems that will emerge. And also... What kind of problems? Well, I mean, socioeconomically, there are, like you said, one of the largest employment centers in the world. I mean, just the economics of cabs would fall apart. But again, we are nowhere, nowhere, nowhere near that. We're not even close. Waymo has had to do the smallest rollouts and the most controlled things because the problem with pretty much every AI system, but especially driving, is not the getting 95% of the way.

[1:02:24](https://realalpha.app/v/Lf5oqGOCRCM/?t=3744) It's those edge cases. It's raining, which is a big problem for them in San Francisco. It's a kid runs across the road, but they're wearing a high-vis thing. Yeah. Does it even notice it's a child? Again, this is a really interesting but very, very applicable example of the right comparison to be made shouldn't be autonomous vehicles versus perfection. It should be autonomous vehicles versus human drivers.

[1:02:45](https://realalpha.app/v/Lf5oqGOCRCM/?t=3765) I mean, I don't know if I agree because the human driver might make mistakes, sure, but... Again, not an expert in autonomous cars. Just want to be clear. But if we're pushing autonomous cars out there willy-nilly, and we're not doing so in extremely controlled environments, those edge cases will multiply and be dangerous. Yeah, they might be better at human drivers in some ways, but they might also...

[1:03:05](https://realalpha.app/v/Lf5oqGOCRCM/?t=3785) I was in Vegas the other day, and I was in a hotel, and I watched a bunch of Zoox cars just get fucking stuck. The autumnalous cars. Yeah, yeah, yeah. These weird boxy things. And they just blocked the exit. They just all kind of lined up and just fell asleep. I saw the same thing actually happen outside of a hotel when it got out of a Waymo in San Francisco. It just stopped at the... And then a bunch of cars and another Waymo got stuck behind it.

[1:03:25](https://realalpha.app/v/Lf5oqGOCRCM/?t=3805) And these are kind of... I've seen some human bad drivers as well. I agree. But it's just we have control over deploying these... Bad or good drivers. We have an ability to roll them out slowly, which is exactly what we should do. I'm not saying autonomous cars are bad. I'm saying we need to be so, so, so careful and treat them as guilty until proven innocent because we can prove... And also, they have people overlooking them.

[1:03:51](https://realalpha.app/v/Lf5oqGOCRCM/?t=3831) They actually have people monitoring the routes. It is something they cannot rush out. And it doesn't seem like they're rushing it, which is good. And they're not promising the world. I do agree. Listen, I'm a big fan of taxi drivers generally, in part because I spend a lot of time in taxis. And I think I'm not just getting in there because I want to get from A to B. I'm getting in there for lots of other reasons.

[1:04:09](https://realalpha.app/v/Lf5oqGOCRCM/?t=3849) However, when I look at the stats... around what is more dangerous, driving myself or having an autonomous vehicle drive me. There's a 68% lower overall crash involvement rate when you're in an autonomous vehicle. Autonomous vehicles experience roughly 2.1 police-reported crashes per million miles compared to humans that are at roughly 4.68 per million miles. So a 55% reduction when you get in an autonomous vehicle.

[1:04:34](https://realalpha.app/v/Lf5oqGOCRCM/?t=3874) Right. An autonomous vehicle show an 80% to 81% reduction in crashes resulting in injuries versus human drivers. So you're 85% less likely to be involved in a single vehicle crash, like hitting a wall or a tree, if you're an autonomous vehicle versus being driven by humans. I agree, but... So it's safer. Also, that data is, what's the sample size of human drivers? I mean, we've got many, many, many, many, many, many more years of drivers and many, many, many more years of accidents.

[1:05:04](https://realalpha.app/v/Lf5oqGOCRCM/?t=3904) And also, man, does that not have anything to do with generative AI. If we were just talking about that... be having a different conversation. I guess the question here was really around job disruption. Like, you know, we look across industries and we go, driving's a massive profession. Is there going to be job disruption because cars can now drive themselves? If we think about white-collar, you know, jobs, you know, lawyers and accountants, people sit here and they tell me that lawyers and accountants, the profession, right, I should say some of the skills within the ai's to do here's the thing lawyers for example great example always hearing fucking legal partners talking about ai never the associates the associates are the ones that go out and find the precedent they're the ones that go and do the grunt work they're the ones who pull emotions half the time the partner is the one that might be the litigant it may be the client facing but the ones that are actually doing the day-to-day work i'm not hearing from them i'm not hearing associates being like this is fucking awesome i'm hearing a bunch of well-paid people that

[1:06:04](https://realalpha.app/v/Lf5oqGOCRCM/?t=3964) Yeah, I'm the greatest lawyer ever. They're not the ones that I want to hear from, the actual workers. White-collar labor disruption is not happening. OpenAI had a study that came out, I think, like a week ago, that said there was no connection between spending on AI tokens and revenue per employee. Like, this is open. What does that mean? Could you explain that to me? As in, the more tokens you spend has no correlation at all.

[1:06:27](https://realalpha.app/v/Lf5oqGOCRCM/?t=3987) with the amount of money you make. It's the second report they've put out. The other one was like, hallucinations are mathematically guaranteed. Kind of almost, it's the one thing I respect about that company. Occasionally, they just put out a study. It's like, yeah, shit, it kind of sucks. But the people that are having their lives disrupted, work-wise, are art directors. It's people, art directors, transcribers, translators, who have bosses that don't care about the output.

[1:06:52](https://realalpha.app/v/Lf5oqGOCRCM/?t=4012) It's what they consider cheap work. And the problem is, is those people would have automated your work away anyway. They would have sold it. They would have taken the cheapest offer. They would have sold it to the global self. They would have taken the shittiest option they could. That is something that AI is doing. And again, those people are not paying the actual cost of AI. They're using a subscription.

[1:07:09](https://realalpha.app/v/Lf5oqGOCRCM/?t=4029) The actual white-collar labor force might have some things that are slightly changing, but there is no evidence of productivity gains. In fact, if there were, they would be screaming it from the rooftops.