Stories of an AI-led “jobs apocalypse” are drastically exaggerated. Or no less than that’s what Sam Altman now claims.
Throughout the Commonwealth Financial institution of Australia convention final week, the OpenAI CEO admitted he could have been unsuitable about some predictions—particularly the velocity with which articial intelligence would feed a considerable chunk of workplace jobs into the digital wooden chipper.
“I’m delighted to be wrong about this,” Altman mentioned. “I believed there would have been extra affect on entry-level white-collar jobs being eradicated by now than has really occurred.”
Anybody nervous about their future employability, nonetheless, may not need to cease refining their résumé simply but. Altman tends to make a number of wild tech predictions, a number of of which he’s already needed to walk back. It’s nonetheless entirely possible that layoffs will ramp up soon, forcing him to stroll again the walk-back.
The issue is that, in contrast to a few of his extra bombastic peers, a few of Altman’s loftier predictions have ended up proving correct. Simply how critically ought to observers take the following forecast, as OpenAI prepares to follow its chief competitor Anthropic in filing an initial public offering?
Maybe one of the simplest ways to determine is by taking an in depth take a look at his observe report to date.
Predictions that proved correct
If Altman has a penchant for grandiosity, he’s no less than considerably entitled. Traditionally, few individuals have lived to see the grandness of their imaginative and prescient so totally realized.
“Within the subsequent 5 years,” he wrote in a 2021 blog post to a pandemic-addled public, “laptop applications that may suppose will read legal documents and give medical advice.”
Whether or not what AI chatbots do can precisely be described as “considering” stays the topic of debate, however Altman was 100% proper concerning the mass adoption of AI that will comply with the discharge of ChatGPT in 2022. Within the years since, it’s grow to be a general-purpose digital software for tens of millions of customers—a Google that may additionally write a guide report, or strains of code, on a consumer’s behalf. (One thing Google itself has since become as properly.)
“Finally, you’ll simply ask the pc for what you want and it’ll do all of those duties for you,” Altman told the audience at OpenAI’s inaugural developer convention in November 2023. He in all probability ought to have pressured the important significance of double-checking AI’s work on these duties, however he was not unsuitable concerning the impending scale and scope of adoption.
Across the similar time, he issued one of his more ominous predictions—that AI would grow to be able to “superhuman persuasion” properly earlier than the arrival of synthetic basic intelligence (AGI) or synthetic superintelligence (ASI).
Sadly, this one has additionally allegedly come to go. The families of several people who’ve died by suicide now claim that ChatGPT assisted in their deaths, with lawsuits pending.
Solely time will inform which different methods AI’s superhuman persuasiveness will take maintain in society.
Predictions that fell quick
A few of Altman’s predictions come throughout as deliberately imprecise. Whether or not advancements in AI capabilities between 2025 and 2027 end up surpassing those between 2023 and 2025, as an illustration, received’t be straightforward to show—which suggests it received’t be straightforward to disprove both.
After all, a few of his different predictions have had quantifiable outcomes that got here and went.
In October 2015, whereas he was nonetheless president of Y Combinator, Altman appeared at Self-importance Truthful‘s New Institution Summit alongside future nemesis Elon Musk. Throughout their time onstage, the tech titans mentioned many subjects, together with the supposedly imminent arrival of self-driving cars, which Altman claimed had been coming “a lot quicker than individuals suppose.” Sadly for Altman, he didn’t keep safely imprecise concerning the timeframe, as a substitute including that self-driving vehicles had been a mere “three to 4 years” away. It finally took almost a decade for fully autonomous vehicles to emerge, and for now no less than they’re carefully geofenced to restricted areas.
Whereas self-driving vehicles nonetheless have a number of kinks to work out, they appear to be a lot additional alongside than synthetic basic intelligence.
Altman first publicly mentioned AI that matches or exceeds human intelligence through the launch of OpenAI in 2015, properly concluding “it’s onerous to foretell” when the tech may be inside attain. By 2024, although, he would consult with it as presumably coming within the “reasonably close-ish future,” solely to later bizarrely floor the thought in a concrete, near-term timeframe the next 12 months.
“We are actually assured we all know how one can construct AGI as we now have historically understood it,” Altman wrote in a January 2025 blog post. “We consider that, in 2025, we might even see the primary AI brokers ‘be part of the workforce’ and materially change the output of firms.”
Agentic AI did in truth hit the workforce in 2025, however the tech stays plagued with reliability issues and falls properly under the bar of AGI.
No marvel Altman pivoted in August to claiming that AGI is “not a super useful term.”
Maybe he’ll have extra luck round his predictions for ASI, which he claimed in a 2024 blog post may be coming within the comfortably distant timeframe of “just a few thousand days.” (Also referred to as 10 “Soras,” a time period I simply invented to explain the mere 10 months OpenAI’s text-to-video mannequin lasted earlier than it was discontinued earlier this 12 months.)
The jury’s nonetheless out on these predictions
As a lot as Individuals have fortunately built-in AI into their lives, they’ve additionally demonstrated a number of combined emotions about it. In keeping with a recent poll from NBC News, as an illustration, 57% say the dangers of AI outweigh its advantages, in contrast with 34% who mentioned the other.
That 57% will not be going to love what the long run has in retailer if Altman’s predictions come true.
In a much-discussed blog post final summer time, the CEO claimed “robots that may construct different robots (and in some sense, knowledge facilities that may construct different knowledge facilities) aren’t that far off.” It would even occur across the time Altman predicts intelligence will grow to be “a utility, like electrical energy or water, and other people purchase it from us on a meter.” Altman hasn’t supplied any predictions, nonetheless, about how Individuals may react when extra of them be taught the enlargement of knowledge facilities is a part of what’s been driving up utility bills currently. (The 70% of Americans who already oppose the development of an AI knowledge middle of their space received’t be happy with these predictions both.)
When AI does surpass human intelligence by 2030, in line with Altman, regardless of some “strange and scary moments,” society received’t change as a lot instantly as one may anticipate. It’s in all probability for the perfect that he hasn’t mentioned far more about simply what can be unusual and scary in these moments, whether or not it should contain environmental calamity or the lack to earn a residing.
As for Altman’s critics within the tech world, properly, they’ve some predictions of their very own.
In keeping with a recent New Yorker piece, a number of colleagues of Altman’s allege he has a penchant for mendacity—together with an unnamed Microsoft exec who predicted there’s a “small however actual likelihood” the OpenAI CEO may finally be remembered as a “Bernie Madoff- or Sam Bankman-Fried-level scammer.”
