AI already does 12% of the work in the US, according to MIT
An MIT study revealed that Artificial Intelligence (AI) already performs 12% of the work in the US, reshaping thousands of jobs in various sectors
More than 30 years ago, the Internet revolutionized communication and access to information. Today, we are experiencing another innovation that will irrevocably change the way we do things: Artificial Intelligence (AI). According to a new study from the Massachusetts Institute of Technology (MIT), this tool already performs 12% of the work in the US. This ignites debate and fears about the reach of this technology in our daily lives. What do you think about this situation?
Artificial intelligence is no longer just a support tool for programmers or data scientists. Today, it impacts administrative, financial, clinical, and logistical processes that form the backbone of the American labor market. The MIT study analyzes how AI is reshaping a labor market valued at $9.4 trillion.
Researchers explain that when an automotive plant incorporates AI in quality control, the effects ripple outward to logistics, supply chains, and even local services that depend on those operations. This occurs because traditional indicators only measure what happens after a disruption.
AI, on the other hand, generates impacts long before companies formalize its adoption. To close this gap, MIT developed Project Iceberg, a model that simulates the interaction between 151 million workers and more than 32,000 skills related to thousands of AI tools. Its main contribution is the Iceberg Index, an indicator that measures the salary value of the skills that AI systems can already perform within each occupation. The authors clarify that this index reflects technical exposure, not layouts or replacement predictions. The analysis reveals that what we see at first glance is just the surface. The study indicates that visible adoption in technology sectors represents “only the tip of the iceberg,” hence the project's name. While exposure in computing is equivalent to 2.2% of the salary value (about $211 billion), the true capacity of AI extends much deeper. The report indicates an 11.7% exposure, around $1.2 trillion, in cognitive tasks across administrative, financial, and professional services. In other words, AI has the capacity to perform nearly 12% of the US workforce.These results also measure the potential for automation of a job or industry. “Financial analysts won't disappear, but AI systems could provide their capabilities in significant parts of document processing and routine analysis,” the researchers stated. “This redefines the structure of jobs and the skills that remain in demand, without necessarily reducing the workforce.” Furthermore, MIT emphasizes that these effects are not concentrated in Silicon Valley or New York. They are distributed across all states. Indicators such as GDP, income, or unemployment explain less than 5% of this skills-based variation. That's why researchers say new metrics are needed to capture the true structure of the AI-driven economy.
An important point is that the study doesn't attempt to measure how many jobs have already been displaced or how many might disappear. Adoption depends on business decisions, public policy, and social acceptance. In some cases, AI complements workers. In others, it completely transforms the labor structure.
Several sectors are already showing clear changes. In healthcare, for example, AI completes administrative paperwork and frees up clinical time for direct care. In manufacturing, it automates quality inspections. In logistics, it streamlines fulfillment and shipping processes. In finance, it classifies documents and supports massive analyses. These advances illustrate that AI is infiltrating routine and cognitive tasks that previously seemed secure. It's important to note that for AI to perform these tasks, it still requires at least one worker to program, receive, or review the data that the AI ????generates.
You may also be interested in:

