How Bad Will AI Jobs Loss Be?

0:00On today's AI Breakdown, we're discussing AI jobs lost, inspired by an exchange between Senator Richard Blumenthal and OpenAI CEO Sam Altman at the Senate's AI hearing on Tuesday. The AI Breakdown is a daily video and podcast about the most important news and discussions in AI. If you're enjoying the AI Breakdown, please like, subscribe, and share.

0:22Is mass job destruction our worst AI nightmare? Welcome back to the AI Breakdown. So there was this super weird moment during the Senate hearing on AI, where the leader of the subcommittee that was holding the hearing asked Sam what his worst nightmare on AI was, but did it in a kind of leading way. You'll see, he quoted something that Sam had said, but then retroframed it in a very strange way. Let's just listen into it. I think you have said, in fact, and I'm going to quote, development of superhuman machine intelligence is probably the greatest threat to the continued existence of humanity, end quote.

1:03You may have had in mind the effect on jobs, which is really my biggest nightmare in the long term. Now, the AI safety folks were just absolutely gobsmacked with that. And in fact, during the hearing, Gary Marcus, one of the other witnesses, actually had to drag it back to get Sam to answer what his actual nightmare around AI was, which is, of course, very much more existential than job loss. However, job destruction is nothing to sneeze at and is something that is big time on the minds of policymakers and individuals as they see the rise of generative AI. This week, we got news that major British telecom company BT is planning to cut around 55 ,000 jobs by 2030.

1:48So over the next seven years or so, and of those 10 ,000 or more are going to be replaced by AI, which is obviously a huge, huge number. And it feels like every week we get some announcement like this. Earlier in the month, IBM announced that they planned to replace 7 ,800 jobs that they were going to hire for with AI instead. These are often back office jobs, human resource jobs, jobs in many cases that don't interface with customers. And these are big numbers, right? Thousands and thousands of jobs at a time. This is part of why the World Economic Forum is estimating that over the next four years, the world will in total lose something like 83 million jobs while it will only gain 69 million jobs, a net loss of 14 million jobs.

2:28And you can see the anxiety over AI-wrought disruption cutting across basically every industry. Music is certainly feeling it, especially as there's this rise in deepfake AI songs. Sting was quoted in a CNN piece recently saying, the building blocks of music belong to us human beings. That's going to be a battle we all have to fight in the next couple of years, defending our human capital against AI. Meanwhile, the Hollywood writer's strike is going on right now, And a big underlying issue is that the studios wouldn't even negotiate with them, wouldn't even share with them their plan as it relates to AI.

3:00The writers are worried, reasonably, I think, that executives might see AI and say, hey, maybe we can get AI-generated scripts and then just hire people to punch them up a little bit. And again, the fact that they wouldn't share their plans or even negotiate around it was something that really, really made the writers nervous. When BuzzFeed started investing in AI, it said that it was really only going to be focused on non-news content. But Brian Merchant, tech columnist for LA Times, pulled out this nugget earlier this week. BuzzFeed CEO Jonah Peretti speaking at Investor Day just one week after laying off its entire news team.

3:31Over the next few years, generative AI will replace the majority of static content. Marketers are, of course, terrified when more than a thousand were surveyed about whether they were concerned about the future discovery of their content. 21 % were extremely concerned. 21 % were moderately concerned. Only 17 % were not at all concerned. And this is to say nothing of search marketers who see in the changes of chat-based interfaces a total difference for how the web has been architected over the last 20 years with major, major implications for how they've designed their businesses. Even coders aren't safe.

4:03In this viral post from Reddit, a developer writes, ChatGPT slowly taking my job away. So I work at a company as an AI ML engineer on a smart replies project. Our team develops machine learning models to understand conversation between a user and its contact and generate multiple smart suggestions for the user to reply with, like the ones that come in Gmail or LinkedIn. Existing models were performing well on this task, while more models were in the pipeline. But with the release of ChatGPT, particularly its API, everything changed. It performed better than our model, quite obvious with the amount of data it was trained on, and is cheap with moderate rate limits.

4:35Seeing its performance, higher management got way too excited and have now put all their faith in ChatGPT API. They are willing to ignore privacy, high response time, unpredictability, and more concerns. They have asked us to discard and dump most of our previous ML models, stop experimenting any new models, and for most of our cases, use the ChatGPT API. Not only my team, but the higher management is planning to replace all machine learning models in our entire software by ChatGPT, effectively rendering all machine learning based teams useless. Now there is low-key talk everywhere in the organization that after integration of ChatGPT API, most of the machine learning-based teams will be disbanded and their team members fired as a cost-cutting measure.

5:12Big layoffs coming soon. The register captures the move in an op-ed that's called, AI is great at one thing, driving the next wave of layoffs. Now there is a flip side to this. Goldman Sachs strategists wrote earlier this week that they see that AI could boost S &P profits by 30 % or more over the next decade. That's driven almost entirely by the productivity gains that AI can produce. For a very long time, the world has not seen significant productivity gains, and there are some macro analysts who think that this might be a hugely important boom for the world once we get through the transition.

5:44The question, of course, is how we get through the transition. Italy announced this week that they were trying to help workers retrain with a new $30 million fund. The Washington Post writes, it will fund companies and nonprofits for projects to train their workers to use new technologies, which can involve robotics, data science, and artificial intelligence. A person involved said, quote, it is necessary to adapt the know-how of workers with training on digital and soft skills so that people can do their jobs in a complementary way. This will enable companies and workers to experience it as an opportunity and not a threat.

6:15Another part of the plan is to put money towards helping people who are currently unemployed or economically inactive to develop new digital skills as well. Now, over in the U.S., you're seeing the exact conversation that Andrew Yang tried to get us to have around AI-based UBI to start to actually get into mainstream media. Lowry writes in The Atlantic this week, before AI takes over, make plans to give everyone money. The U.S. needs policies now to support workers made redundant by artificial intelligence. This article also quotes Goldman Sachs economists, and these ones suggest that in the coming decade, AI could wipe out 300 million jobs, one out of every 11 jobs on the planet.

6:52Now, when Sam Altman was asked about this at the hearing, he said a thing that he said before that he genuinely believes that the jobs of future will be better than the jobs of today, that what AI will take away is the menial work, and that people will find new, much better things to spend their time on. With all technological revolutions, I expect there to be significant impact on jobs, but exactly what that impact looks like is very difficult to predict. If we went back to the other side of a previous technological revolution, talking about the jobs that exist on the other side. You know, you can go back and read books of this.

7:27It's what people said at the time. It's difficult. I believe that there will be far greater jobs on the other side of this, and that the jobs of today will get better. I think it's important. First of all, I think it's important to understand and think about GPT-4 as a tool, not a creature, which is easy to get confused, and it's a tool that people have a great deal of control over in how they use it. And second, GPT-4 and things, other systems like it, are good at doing tasks, not jobs. And so you see already people that are using GPT-4 to do their job much more efficiently by helping them with tasks.

8:05Now, GPT-4 will, I think, entirely automate away some jobs. And it will create new ones that we believe will be much better. This happens, again, my understanding of the history of technology is one long technological revolution, not a bunch of different ones put together. But this has been continually happening. As our quality of life raises and as machines and tools that we create can help us live better lives, the bar raises for what we do. And our human ability and what we spend our time going after goes after more ambitious, more satisfying projects. So there will be an impact on jobs. We try to be very clear about that.

8:44And I think it will require partnership between the industry and government, but mostly action by government to figure out how we want to mitigate that. But I'm very optimistic about how great the jobs of the future will be. The question is whether it's his view of the future, that view of the future, that's correct. Or if it's this view here from Lauren Marie on Twitter, who writes, My 3am thought, how does our current economic model, which compensates people based on their alleged labor value to society, continue to exist in the future when AI makes 97 % of jobs obsolete? I think about this every day.

9:16Tech leaders and legislators aren't even really addressing this publicly yet. I don't know the answer, but now is the time to have the conversation. That's it for today's AI Breakdown. If you're enjoying the show, please like, share, and subscribe. Go check out the podcast and the newsletter. And until next time, peace.

9:35Thank you.