In a week already full of tense debates about artificial intelligence, a powerful new report from the Massachusetts Institute of Technology has landed like a cold splash of water on the U.S. labour market. According to a brand-new study released on November 26, 2025, AI systems available today—not in some far-off future—could technically take over 11.7% of all U.S. jobs.
That number represents nearly $1.2 trillion in wages, and the findings are stirring deep concern among economists, tech leaders, and everyday workers who now fear the ground shifting under their feet.
MIT researchers, working with the Oak Ridge National Laboratory (ORNL), have created something they call the Iceberg Index—a massive digital simulation that models more than 151 million American workers, nearly 900 occupations, and over 32,000 work-related skills. The name “Iceberg” is intentional. The tiny part visible above the surface—the small share of AI-impacted tech jobs we talk about daily—is nothing compared to the huge, hidden mass of white-collar jobs that are exposed below the waterline.
It’s a finding that many experts say should serve as a wake-up call.
A Digital Twin of the U.S. Workforce
“Basically, we are creating a digital twin for the U.S. labour market,” explained Prasanna Balaprakash, ORNL director and co-leader of the research, in comments to CNBC.
The system is capable of running large-scale “what-if” experiments on how AI might shift tasks, skills, and movement inside the workforce—years before those changes show up in real government economic reports.
The Iceberg Index doesn’t come with flashing warnings about which companies will cut staff next year. It isn’t trying to predict exact job losses. Instead, it measures one thing with scientific precision: what AI is technically capable of doing today.
The number: 11.7% exposure.
To put it plainly, nearly one out of every nine American jobs contains enough tasks that today’s AI could handle them at human level—or better.
And it’s not just Silicon Valley that’s vulnerable. MIT researchers say the impact reaches deep into sectors once considered “safe” from automation:
- Finance and accounting
- Healthcare administration
- Professional services
- Customer support
- Legal and compliance tasks
- Back-office operations
- Administrative and clerical roles
Tech jobs—ironically—represent only about 2% of the exposed wage value. It’s the white-collar, cognitive work that is the real iceberg under the water.
A Snapshot of Today — Not 2030
One of the most misunderstood parts of the report is timing. The Iceberg Index does not say that 12% of jobs will disappear next year. Instead, it shows what AI could do today if companies decided to adopt those tools aggressively.
In simple terms:
The technical capacity is already here. The adoption timeline is what remains uncertain.
This distinction between task exposure and actual job displacement is becoming a central point in the national debate.
Still, for workers reading the headline—“AI can replace 12% of jobs today”—the reassurance may feel thin.
Why This Study Is Different — And Why It’s Causing a Stir
For years, predictions about AI wiping out jobs felt like something between science fiction and distant-future economics. But the MIT–ORNL model is the first to use real, current AI capabilities to simulate the entire labour market at task level.
That means the numbers here aren’t guesses. They’re grounded in what the latest generation of AI—including generative models—can actually do.
This is why business outlets and tech journalists quickly started drawing comparisons between the Iceberg Index and the rising anxiety coming from AI leaders themselves.
And those warnings, frankly, have been getting sharper.
Tech Leaders Worried: “We Have a Duty to Tell the Truth”
Back in May, Anthropic CEO Dario Amodei delivered one of the bluntest warnings the industry has seen. He suggested that AI might wipe out up to 50% of entry-level white-collar jobs within five years, pushing unemployment into territory the U.S. hasn’t seen in generations.
“We, as the producers of this technology, have a duty and an obligation to be honest about what is coming,” Amodei said. “I don’t think this is on people’s radar.”
According to Amodei, workers and governments are dangerously unprepared. He has compared the coming shift to a “slow and silent earthquake” that will hit admin staff, junior analysts, paralegals, HR assistants, junior finance workers, and other early-career office roles first.
The MIT study does not directly predict these outcomes, but its findings give real weight to Amodei’s worries.
If AI can already handle 12% of tasks nationwide, it’s not hard to imagine companies moving faster than policymakers expect.
Geoffrey Hinton: “It Will Make a Few People Richer and Most People Poorer”
Meanwhile, Geoffrey Hinton—often called the “godfather of AI”—continues to say the quiet part out loud.
Hinton has been warning for over a year that AI could sharply increase inequality. While companies enjoy record efficiency and profits, workers may be pushed aside in a way modern economies have not seen since the Industrial Revolution.
“What’s actually going to happen is rich people are going to use AI to replace workers,” Hinton told the Financial Times.
“It’s going to create massive unemployment and a huge rise in profits. It will make a few people much richer and most people poorer. That’s not AI’s fault, that is the capitalist system.”
Hinton’s blunt critique echoes the deepest fear emerging from the Iceberg Index: AI’s benefits will not be evenly shared.
Unless, he says, governments intervene.
Why This Matters for Workers Right Now
The MIT researchers behind the Iceberg Index stress that the model is meant to help policymakers make better decisions before real jobs are lost. It’s a tool to test:
- What if AI spreads faster in finance?
- What if companies automate administrative work first?
- What if reskilling programs focus on healthcare?
- What if regulations slow down certain types of automation?
Think of it as a warning system—one that lets policymakers “see around corners” instead of reacting after millions of workers have already been displaced.
But for workers, the message feels more immediate than academic:
AI is not coming for jobs someday.
AI is capable of taking jobs right now.
Whether companies will actually pull the trigger this year or next is the unknown part.
The Bigger Debate Is Shifting: It’s No Longer “If” — It’s “How Fast”
Across corporate America, something subtle but important is changing.
For years, boardrooms and executives asked whether AI would matter to their operations. Whether it was more hype than substance. Whether investing heavily made sense.
The Iceberg Index suggests that era is over.
The debate is now about speed:
- How fast will companies redesign roles?
- How fast will workers need new skills?
- How fast can regulators respond?
- How fast will inequality grow if nothing changes?
Because AI’s technical capabilities are no longer theoretical—they’re measurable, and the numbers are large.
The Cold Reality: This Is Just the Beginning
One sobering detail from the MIT report: the 11.7% estimate is based on current AI, not future upgrades.
And the researchers admit openly that these systems are improving quickly.
If today’s tools can technically do the work of 12% of the U.S. labour market…
What will next year’s tools be capable of?
And the year after that?
Amodei warns that entry-level white-collar work could be devastated in five years. Hinton says inequality will skyrocket without major intervention. Economists fear that unemployment could rise enough to pressure the entire safety-net system.
Suddenly, the MIT study isn’t just another academic report. It’s a flashing warning sign.
A Nation at a Crossroads
The U.S. now faces a choice, according to the study’s authors and many experts who reviewed the findings.
The choice is not whether AI will change work.
The choice is whether the country will prepare for it—or be blindsided by it.
For policymakers, the Iceberg Index is a rare opportunity to get ahead of a technological wave instead of scrambling behind it. For workers, the study may be the clearest signal yet that training, upskilling, and career flexibility are becoming non-negotiable.
For companies, it raises an uncomfortable question:
Will they use AI to empower their people…
or replace them?
Because the truth is simple:
AI isn’t waiting. It’s already capable. It’s already here.
And the iceberg is bigger than anyone thought.
Login with Google