If AI Can Do 12% of US Jobs, Why Layoffs Haven’t Begun Yet: MIT Report Explains

Artificial intelligence is now capable of performing tasks equal to nearly 12% of all job functions across the US economy, according to new research from MIT’s Project Iceberg. But despite this significant capability, the US is not witnessing widespread layoffs or workforce displacement.

Instead, companies are quietly redesigning job roles, particularly at the entry level, rather than reducing staff.

AI Can Replace Tasks — Not Whole Jobs

The MIT study, which used data from 151 million workers and more than 32,000 job skills, found that AI excels in repetitive, pattern-based tasks such as data processing, document checks, reconciliation, reporting and routine analysis.

But experts stress that this does not translate to eliminating entire roles.

“AI is improving speed and accuracy, but people still handle judgment, interpretation and oversight,” said Sripal Jain, CA/US CPA and Co-founder of Simandhar Education.

This explains the absence of mass layoffs:
AI is automating tasks, not jobs.

Why Entry-Level Roles Are the First to Shrink

Beginners are most affected because their jobs largely involve the very tasks AI now performs efficiently. As a result:

• Hiring of freshers has slowed

• Job descriptions are changing

• Companies expect new hires to work with AI, not do manual tasks AI can already handle

“Companies now want people who can interpret AI’s outputs, not just execute repetitive steps,” Jain added.

AI literacy, reasoning, and domain expertise are rapidly becoming essential.

AI Adoption Is Slow Because Real-World Integration Is Difficult

Even if AI can theoretically perform 12% of tasks, most organisations cannot implement it overnight.

SP Jain School of Global Management Dean Dr. Balakrishna Grandhi explained that AI adoption requires:

• Training staff

• Workflow redesign

• Quality checks and supervision

• Higher upskilling investments

Companies prefer to upskill existing employees rather than replace them, which slows job cuts but increases skill expectations.

A Silent Transformation: Workflows Are Changing, Not Workforces

Although only 2.2% of the US wage value reflects active AI usage today, AI is technically capable of affecting up to 11.7% of wages—five times more.

The biggest changes are happening behind the scenes in:

• Finance and audit

• HR and payroll

• Customer support

• Back-office operations

• Logistics and compliance

• Healthcare administration

These sectors are experiencing task-level automation, which rarely triggers headlines but significantly alters operations.

The New Skill Economy

AI is reshaping how careers begin, rather than eliminating careers altogether.

Workers who excel in:

• Problem-solving

• Reasoning

• Communication

• AI-assisted tools (Copilot, Power BI, analytics, LLM systems)

will have a competitive advantage.

Domain expertise is becoming even more valuable, since human oversight remains central to AI-driven workflows.

AI Isn’t Taking Jobs — It’s Redesigning Them

The MIT findings suggest the world is entering a new phase of labour evolution marked by:

• Fewer entry-level opportunities

• Higher skill thresholds

• Increased reliance on automation

• Human-AI collaboration

• Continuous upskilling requirements

AI is not creating mass unemployment.
It is redistributing opportunity toward those prepared for the next generation of work.