Sam Altman stood at the India AI Impact summit on February 23, 2026, and called concerns about AI’s water consumption “completely untrue, totally insane.” Fake, essentially.
Meanwhile, January 2026 saw 108,435 US job cuts—the highest monthly total since the 2009 financial crisis. AI was cited in just 7,600 of them. That’s 7%. The other 93%? Those companies blamed everything but the technology everyone’s panicking about.
This is the industry’s selective honesty problem in a single week: dismiss documented environmental costs while inflating undocumented job displacement. And it’s training workers to misread the market at exactly the wrong moment.
The 93% nobody’s talking about: when “AI layoffs” aren’t about AI at all
That 108,435 figure isn’t just big—it’s the highest monthly US job cut total in 17 years. Yet only 7,600 cuts blamed AI. This is AI washing at industrial scale.
Companies use AI as cover for overhiring corrections from 2021-2022 and mundane cost optimization. It’s convenient. It’s forward-looking. It absolves bad hiring decisions.
But here’s Altman at that same summit, equating AI energy consumption to human needs. “We’re all consuming energy,” he said, as if a data center and a person are morally equivalent resource users.
Researcher Matt Stoller wrote on X: “He’s saying a really big spreadsheet and a baby are morally equivalent.” Another user: “Anybody who talks like this about humans should not be allowed a job that impacts other humans.”
The contradiction matters because it reveals the pattern. A January 2026 report from Xylem and Global Water Intelligence projects data center water demand will triple over the next 25 years, driven primarily by AI workloads.
That’s documented. Measurable. Altman calls it fake. But AI’s job impact—the thing inflating 93% of January’s layoff narratives—gets admitted when it’s useful. Fortune reported February 19 that Altman himself confirmed “AI washing is real,” acknowledging companies falsely blame AI for cuts they’d do anyway.
Admit the convenient problem. Deny the inconvenient one.
The math problem: 90% of executives see no AI job impact, yet everyone’s blaming AI
A February 2026 NBER study found 90% of C-suite executives report zero AI employment impact from 2022-2025. Zero. Yet AI washing accelerated through that same window. All of 2025 saw 55,000 AI-attributed layoffs—less than 1% of total US job losses, per Oxford Economics.
The danger isn’t the lie. It’s the signal distortion.
Job seekers see “AI displacement” headlines, pivot to “AI-proof” careers, and miss that 93% of January cuts were mundane cost optimization dressed up as technological inevitability. And while that’s happening, IBM is reportedly tripling US entry-level hiring in 2026, suggesting AI is creating demand in some sectors. But AI washing drowns that signal. Workers can’t distinguish real displacement from corporate theater, so they’re making career decisions based on false information.
This is worse than actual AI displacement. At least with real automation, the market signal is honest—these skills are obsolete, learn these instead. AI washing corrupts the feedback loop. It delays real reskilling while companies dodge accountability for hiring 40% too many people in 2021.
The one company actually cutting for AI—and why it’s different
Pinterest announced in January 2026 it’s cutting 15% of its workforce, with funds redirected to AI initiatives. This is genuine AI displacement—cutting humans to fund automation. It’s also rare. Most “AI layoffs” are overhiring corrections with AI branding. But we can’t distinguish in real-time.
Workers see Pinterest plus 7,600 January “AI cuts” and assume a displacement wave. Reality: 100,000+ January cuts had zero AI connection. This misjudgment delays the hard work—figuring out what skills survive genuine AI displacement—while executives avoid owning bad 2021-2022 hiring decisions.
And that brings us back to Altman’s split position. On February 19, he told Fortune that AI washing is real. On February 23, he told the India summit that water use concerns are fake. Zoho CEO Sridhar Vembu, sitting at that same summit, responded: “I do not want to see a world where we equate a piece of technology to a human being.”
If industry leaders only admit problems when convenient, how do workers know which signals to trust?
