An MIT study claims AI now has the capability to replace about 11% of U.S. jobs

MIT puts a hard number on the current AI job risk

A new MIT study suggests that today’s AI systems are already capable of performing work equivalent to about 11.7% of US jobs.

That translates to roughly $1.2 trillion in wages across finance, healthcare, logistics, and professional services. This is not a sci-fi projection decades in the future; it is a snapshot of what existing tools can technically and economically handle right now.

The Iceberg Index creates a digital twin of the job market

To reach that figure, MIT and Oak Ridge built the Iceberg Index, a massive labor simulation that treats 151 million workers as individual agents.

Each “worker” is tagged with skills, tasks, occupation, and location, spanning over 32,000 skills, 900+ occupations, and 3,000 counties, and then matched against what current AI systems can actually do today.

The key idea is about capability, not certainty

The headline number is about technical and economic capability, not a countdown to guaranteed layoffs. The researchers emphasize that whether jobs are actually automated depends on factors such as adoption costs, company strategy, regulation, and social pushback.

AI can handle a chunk of your role on paper, but firms still have to decide if it is worth the risk, retraining, and change management.

Visible tech layoffs are only the tip of the iceberg

Currently, the most visible impact is in tech and computing roles, particularly in coding. However, the study finds that those headline disruptions represent only about 2.2 percent of total wage exposure, roughly $211 billion.

Beneath that small tip sits a much larger mass of exposed work in quieter back-office roles that rarely grab attention but keep organizations running.

Hidden exposure lies in routine white collar tasks

Where does the rest of the $1.2 trillion sit? Mostly in routine cognitive work. Think of human resources, logistics coordination, finance operations, office administration, and basic document-heavy professional services.

These jobs are full of structured tasks that today’s AI can already read, summarize, classify, and respond to, often more efficiently and cost-effectively than an entry-level analyst or assistant.

Entry level knowledge jobs feel the pressure first

The study notes that AI is already eating into work that traditionally went to fresh graduates. Systems now generate massive amounts of code daily, draft reports, and perform routine analysis.

That does not mean senior engineers or analysts vanish overnight. Still, it does mean fewer pure junior “grind” roles and more pressure on newcomers to arrive with broader skills than just basic execution.

Many jobs will be reshaped rather than erased

MIT’s team is clear that “replace” does not always mean eliminating roles. In many cases, AI automates specific tasks within a job, such as document processing, data cleanup, or form filling.

Nurses might spend less time on paperwork and more time with patients. Financial analysts may offload routine number crunching to models and focus on judgment, communication, and strategy instead of raw spreadsheet labor.

Physical and hands on work still has some insulation

Sectors that lean heavily on physical tasks, such as healthcare at the bedside, manufacturing on the line, transportation, nuclear energy, and field maintenance, appear less exposed in the near term.

The Iceberg simulations show that purely digital tasks are most manageable to automate first. The bigger challenge for these industries is using AI and robotics to augment their workforce without undermining safety or eroding expertise.

The impact is not just a story about a coastal tech hub

One significant myth this study dispels is that AI risk is confined to Silicon Valley or the New York finance sector. By modeling skills across every county, MIT finds exposure spread across all fifty states, including inland and rural regions.

Roles in back-office services, logistics, and administration are prevalent in almost every local economy, so the disruption map does not follow traditional tech cluster lines.

States are already using the model to plan policy moves

This is not just academic. Tennessee has already cited the Iceberg Index in its AI Workforce Action Plan, and North Carolina and Utah are running their own simulations.

Policymakers can drill down to the county or even the census block level, identify which skills are most at risk, and test what happens if they shift training dollars or tweak incentives before committing real money.

The tool works as a sandbox for what-if scenarios

MIT positions Iceberg as a policy and business sandbox, not a crystal ball. Leaders can model what happens when AI adoption accelerates in a specific sector, or when reskilling programs successfully transition workers into complementary roles.

That ability to test different futures on a “digital twin” of the labor market is meant to reduce blind spots and nasty surprises later.

Both workers and companies need to rethink their skills

For workers, the study serves as a nudge to lean into skills that AI struggles with, such as creativity, cross-domain problem-solving, people leadership, ethics, negotiation, and hands-on craft.

For companies, it is a warning that pretending AI is a distant trend is no longer a viable option. The organizations that do best will redesign jobs, not just eliminate them, and invest in serious retraining.

And if you’re thinking about how these shifts play out across whole careers, you might want to see how AI is reshaping the entire job ladder, not just entry-level roles.

The real question is how we choose to respond together

MIT’s headline number sounds scary, but it is really a call for grown-up planning rather than panic. AI can already perform a significant portion of the country’s work, and that share is expected to continue growing.

Whether that turns into mass displacement or a messy but manageable transition depends on what employers, educators, and policymakers decide to do with this warning shot.

And if you’re watching how this shift is hitting the companies building AI, you might want to see why Meta just cut 600 AI jobs amid a major restructuring of its own AI division.

What do you think about MIT studies claiming that AI now has the capabilities to replace about 11.7% of jobs in the US? Please share your thoughts and drop a comment.

Read More From This Brand:

• Gen Z is shrugging off the AI jobs apocalypse, and classrooms prove it
• OpenAI is rolling out an AI jobs platform to support workers displaced by tech
• Game over for jobs? Musk thinks AI rules

Don’t forget to follow us for more exclusive content on MSN.

If you liked this story, you’ll LOVE our FREE emails. Join today and be the first to get stories like this one.

This slideshow was made with AI assistance and human editing.