Major AI tech giants have recently shifted their stance, moving from the “job apocalypse” narrative to the “productivity miracle” story. OpenAI’s CEO Altman publicly retracted his earlier prediction that white-collar jobs would vanish en masse; Anthropic’s CEO Amodei now emphasizes that AI can boost human productivity by up to tenfold; and Musk has suggested that future work will resemble personal hobbies. This shift is driven by multiple pressures: OpenAI and Anthropic are racing toward IPOs and need compelling narratives; public sentiment toward AI has turned sharply negative; actual employment data shows no signs of an apocalypse; and AI’s real costs and performance fall far short of earlier hype. Data shows that companies heavily adopting AI have seen a 10.2% growth in total employee numbers, yet a growing divide has emerged between cutting-edge firms and traditional enterprises.

Article author and source: Tencent Technology

How does public negative sentiment influence the narrative shift among tech leaders?

Among the AI giants, doomsday theories are outdated, and optimism is making a comeback.

A year ago, OpenAI CEO Sam Altman was publicly warning that many jobs would "completely disappear."

Anthropic CEO Dario Amodei asserted that within five years, half of entry-level white-collar jobs will cease to exist, potentially causing unemployment to surge to 20%.

Palantir CEO Alex Karp has even said that in the AI era, only artisans and those who naturally think outside the box can ensure they are not left behind.

At the time, everyone from Elon Musk to the leaders of traditional automakers like Ford painted a bleak picture of AI decimating office white-collar jobs.

However, this wave of "job apocalypse" theory has recently come to a sudden halt.

Ultraman readily admitted his mistake. Amodei rephrased it, no longer talking about job loss but rather about human productivity being increased tenfold by AI. Musk simply stated that work in the future won't be necessary—it will be more like a personal hobby.

They suddenly changed their stance due to multiple pressures.

Because a compelling story is needed for the IPO push, public sentiment against AI has become impossible to ignore, employment data shows no signs of a "doomsday" scenario, and the actual costs and performance of AI fall far short of the initial hype.

01 From "Doomsday Prophecy" to "Productivity Myth"

At the end of May 2026, Otomo publicly reflected at an industry conference in Sydney, acknowledging that the industry had underestimated the core value of “people” in economic interactions. He admitted that the anticipated large-scale disappearance of white-collar jobs had not occurred, and his intuition regarding AI’s impact on the economy was incorrect.

Amody also revised his aggressive stance, instead defining AI as a "productivity multiplier."

In May this year, he stated alongside JPMorgan Chase CEO Jamie Dimon that even if 90% of workflows are automated, the remaining 10% would create new labor demands, potentially increasing individual productivity several-fold.

In a June article, Amodei explained the reasoning behind his changed stance: his initial warnings were intended to help policymakers prepare better, and he did not aim to be a "doomsday prophet," though he still left room for concern, noting that the risk of persistent unemployment remains.

Ford’s actions reflect this shift. Last year, Ford CEO Jim Farley predicted that AI would replace nearly half of all white-collar jobs in the U.S., but recently the company has instead hired hundreds of additional engineers, citing the need for highly skilled engineers to oversee the quality of automation tools.

Goldman Sachs CEO David Solomon, from a historical cycle perspective, noted that every major technological disruption in U.S. history—from electrification to the digital revolution—has been accompanied by the emergence of new employment ecosystems. The firm’s research shows that AI-driven data center construction alone has created 200,000 jobs since 2022.

Research by Nobel laureate Daron Acemoglu also confirms that the displacement effects of AI are typically offset by new labor demands driven by productivity gains.

The fintech company Ramp, in collaboration with the workforce intelligence company Revelio Labs, tracked AI investment and hiring data from nearly 22,000 U.S. companies.

The report shows that organizations classified as "high-intensity adopters" in AI—those spending over $30 per employee per month on AI—experienced an employee growth rate of 10.2%, with increases across roles in engineering, sales, administration, finance, and more.

This phenomenon validates the economic principle known as the Jevons paradox, which states that when technological progress increases resource efficiency, overall consumption tends to rise rather than decline.

Box CEO Aaron Levie and Apollo’s Torsten Slok both noted that AI has reduced the unit cost of core outputs such as coding and customer interaction, thereby stimulating companies to expand their business boundaries and increasing overall labor demand.

Another set of macro data from Goldman Sachs shows that over the past year, AI has net eliminated approximately 16,000 jobs per month, with Gen Z and entry-level employees bearing the brunt. However, at leading companies at the forefront of technology, the number of entry-level employees has actually increased by 12%.

This subtle contradiction reveals a harsh reality: AI is creating polarization. Tech companies at the forefront are rapidly expanding their teams, while most traditional businesses trapped in experimentation and lacking sustained investment are suffering the greatest job losses.

02 Preparing for the IPO

Given how complex and varied the impact on the job market is, why have these tech leaders changed their statements so quickly?

Currently, OpenAI is preparing to secretly file its initial public offering (IPO) application, targeting a valuation of $1 trillion and planning to raise at least $60 billion, with the goal of achieving $280 billion in revenue by 2030. Meanwhile, Anthropic has also submitted a confidential S-1 filing, with its valuation approaching the $1 trillion mark.

AI strategy consultant Bob Hutchins noted that companies cannot gain the trust of bankers and retail investors in public markets by invoking narratives of societal collapse and mass unemployment. In anticipation of upcoming regulatory scrutiny and IPO fundraising, major corporations must adjust societal expectations.

He explained that in 2025, CEOs made bold statements to tech media, when such remarks were popular. But by 2026, their audience had shifted to bankers, retail investors, and ordinary people who had grown tired of the hype. The audience no longer bought into it, so the messaging had to change accordingly.

Additionally, negative public sentiment toward AI is mounting.

An NBC poll shows that AI's net positive rating has fallen into negative territory. Gallup surveys also indicate that younger generations are experiencing growing anxiety and resistance toward AI, even sparking offline protests against data center construction and tech executives.

Even if the warnings about unemployment are well-intentioned, they directly clash with a population deeply troubled by employment anxiety.

When news emerged about ChatGPT’s release and claims that jobs would be replaced, it coincided with the tech industry undergoing massive layoffs after years of over-hiring. Further warnings of job losses from executives struck a nerve with workers already worn down by uncertainty. These claims conveniently became the perfect justification for corporate layoffs—job cuts framed as an inevitable response to the tide of technological advancement.

MIT economics professor David Autor bluntly stated that tech giants realize it’s a terrible business move to claim their groundbreaking new products will destroy the social economy. In promoting data center construction and navigating government regulations, downplaying fears of job losses carries a deliberate political intent.

ROI (return on investment) concerns at the business level are also compelling companies to return to rationality. Emergn’s survey shows that most U.S. business leaders struggle to see tangible returns on their AI investments.

Meanwhile, the high cost of computational power has created a bottleneck for technology adoption. Bryan Catanzaro, Vice President of Applied Deep Learning at NVIDIA, revealed that in certain projects, “computational costs have far exceeded personnel costs.” Giants such as Uber and Microsoft have begun restricting or canceling engineer access to certain AI tools due to rapid budget consumption.

03 AI has become an accomplice in layoffs

Despite a shift toward optimism in the narratives of industry leaders, layoffs in the technology sector continue.

In the first five months of 2026, over 115,000 layoffs occurred in the technology sector. According to Challenger, Gray & Christmas, AI was cited as the reason for nearly 50,000 of these job cuts.

Andy Challenger, a workplace expert at Challenger, Gray & Christmas, says that the essence of layoffs is not that jobs are being fully replaced by AI, but rather a reallocation of corporate funding—budgets previously allocated for human salaries are being redirected toward purchasing computing power and servers.

Notably, about half of the companies that previously cut customer service roles citing AI have since planned to rehire human staff due to quality issues with automation. This demonstrates that AI’s current ability to replace human workers has been significantly overestimated.

Discussions about AI's impact on employment have swung dramatically over the past four years—from an “efficiency myth” to a “job apocalypse”—and now back toward rationality.

In this round of narrative refinement, the most specific footnote came from an experiment conducted by Ultraman himself. Ultraman once tried using an AI agent to respond to everyday Slack messages and emails, but ultimately abandoned it due to the lack of genuine human qualities and emotional connection, choosing instead to return to human responses.

This minor episode in technological history shows that, no matter how algorithms evolve, the core of business society and economic collaboration remains “human-to-human interaction.” Trust, intuition, and emotional resonance among humans in complex commercial environments are still barriers that cold, impersonal code cannot overcome.

This article is from "Tencent Technology," authored by Boyang, edited by Xu Qingyang, and published with authorization from 36Kr.