OpenAI CEO Sam Altman has done something few Silicon Valley bosses ever do, admit he is wrong. Speaking virtually at a Commonwealth Bank of Australia conference in Sydney in May, Altman confessed that one of his biggest concerns about AI simply has not played out the way he expected. For someone whose job often involves predicting the future, it was a surprisingly candid moment.
"I'm delighted to be wrong about this. I thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened," Altman said. "I now think I understand more about why it hasn't, and I'm obviously grateful, but that is an area where my intuitions were just off."
Altman explained that OpenAI had been "roughly right" about many of the technological predictions it made when ChatGPT launched. AI has become more capable at an astonishing pace. What he appears to have misjudged was how those capabilities would translate into changes in everyday employment.
Personal AI experiments
Notably, Altman concluded he had been wrong after an experiment in which he let AI handle some of his own communications. He didn't need a labor market research report to see that it wasn't up to snuff. He used AI to answer Slack messages and emails, each labeled as coming from "Sam's AI" rather than from him directly.
But Altman found himself pulling back from the experiment almost immediately. The reason had little to do with the quality of the responses. Rather, Altman simply didn't want to give up interacting with people to an AI model, no matter how efficient.
"We really do care about our interactions with people and this thing, which is a huge amount of my time, is not something that I can imagine myself outsourcing to an AI anytime soon," Altman said.
The experience appears to have shifted Altman's thinking about employment more broadly. Jobs often look simple when reduced to a list of tasks. In reality, many roles involve trust, relationships, judgment, and personal interactions that are difficult to capture in a spreadsheet.
Human jobs
None of this means Altman suddenly believes AI will leave the workforce untouched. OpenAI continues to release increasingly powerful models, and businesses continue searching for ways to use them more effectively.
But the actual disruption of employment will be less catastrophic, according to Altman. Discussions about AI often treat jobs as collections of tasks that could be swapped out with the right AI prompt, but reality appears messier. Companies may automate parts of jobs long before they eliminate entire positions.
"It really, in both positive and negative ways, updated me to thinking that the jobs picture is likely to be very different than we thought. I don't think we're going to have the kind of jobs apocalypse that some of the companies in our space advocate or talk about."
That distinction matters because it helps explain why the labor market has not experienced the immediate shock that many observers expected. AI has certainly changed a lot of research and enterprise projects. But most organizations still need people to make decisions, manage relationships, and take responsibility when things go wrong.
Altman's more positive view of AI on job prospects doesn't mean there's no problem with how the technology is being deployed. But people who might look to Altman for insight into AI might feel a little better, even if it's just him saying AI will have a muddled influence and not act as a straight assassin of careers.
