AI COULD AUTOMATE 93% OF U.S. JOB TASKS A new study finds AI could handle parts of 93% of U.S. jobs, potentially shifting $4.5 trillion in labor costs. Researchers analyzed 18,000+ tasks across 1,000 jobs, with software development, finance, management, legal, and office roles most affected. Cognizant CTO Babak Hodjat notes adoption is uneven, but rapid AI breakthroughs—like multimodal and agentic AI—are accelerating automation. Some physical and care jobs, like construction and healthcare support, see smaller but growing AI impact. Impact doesn’t mean job loss: AI augments human work, improving efficiency and output. Globally, AI could influence $15 trillion in labor value, with its capabilities expanding fast.
From the Future of Jobs Report 2025, this infographic visualizes the human–machine frontier: the share of work tasks done by people vs. technology now and by 2030. It illustrates AI’s tasks-based impact and growing automation/augmentation, directly aligning with the idea that AI affects many job tasks without necessarily eliminating entire jobs.
Source: World Economic Forum
Research Brief
What our analysis found
A sweeping new report from Cognizant, titled "New Work, New World 2026" and released on January 15, 2026, claims that current AI technology could assist or automate portions of 93% of U.S. jobs, theoretically exposing up to $4.5 trillion in U.S. labor value to AI-driven transformation. Researchers re-scored more than 18,000 tasks across roughly 1,000 occupations using O*NET data and found that AI exposure has accelerated dramatically — average exposure scores are now 30% higher than the firm's own 2032 forecast, and annual exposure growth surged from 2% to 9% compared to a prior study. The share of jobs at the highest exposure level jumped from 0% in 2023 to 30% in 2026, while the lowest-exposure share plummeted from 31% to just 7%.
White-collar professions saw the sharpest spikes. Business and financial operations, management, and office/admin support roles leapt from 14–21% exposure in 2023 to 60–68% in 2026. Financial managers now register 84% exposure, CEO roles exceed 60%, and legal occupations surged from 9% to 63%. Even healthcare practitioners climbed from 33% to 59%. Physical and care-based jobs saw smaller but notable increases — construction and extraction rose from 4% to 12%, while healthcare support jumped from 5% to 29%. Cognizant CTO Babak Hodjat attributes the acceleration to breakthroughs in multimodal models, advanced reasoning, and agentic AI systems capable of planning and acting across tools and workflows.
However, the report's authors and independent analysts caution that these figures represent a theoretical maximum, not a forecast of imminent job loss. Cognizant itself stresses that exposure scores do not account for enterprise adoption rates, regulation, quality control, or organizational inertia. Other major studies — from the IMF, OECD, OpenAI, and MIT — have produced notably lower estimates of AI's reach, and real-world usage data from Anthropic suggests AI is currently used more for augmentation than full automation. Globally, Hodjat has suggested AI could influence roughly $15 trillion in labor value, but the gap between what AI can theoretically do and what economies will actually adopt remains vast.
Fact Check
Evidence from both sides
Supporting Evidence
Cognizant's own report confirms the 93% and $4.5 trillion figures
The "New Work, New World 2026" report landing page and full PDF explicitly state that 93% of U.S. jobs are impacted and approximately $4.5 trillion in U.S. labor value is theoretically exposable, based on a re-scoring of 18,000+ O*NET tasks across ~1,000 occupations (cognizant.com).
Press release and investor materials mirror the findings
A PR Newswire release dated January 15, 2026 reiterates the 93% and $4.5 trillion claims and provides specific examples, such as transportation exposure rising from 6% to 25% and construction from 4% to 12% (prnewswire.com).
Forbes interview with Cognizant AI chief corroborates details
A February 25, 2026 Forbes TechFirst interview with Babak Hodjat repeats the 93% and $4.5 trillion findings, highlights white-collar roles as most exposed, and extends the analysis globally to suggest roughly $15 trillion in labor value could be influenced (forbes.com).
Hodjat's public statements are consistent
LinkedIn posts from Hodjat in early 2026 restate the core claims — "93% of jobs now impacted" and "$4.5T" in U.S. labor tasks — maintaining consistency across channels (linkedin.com).
Acceleration data is internally documented
The report shows average exposure scores 30% higher than Cognizant's prior 2032 forecast, with annual exposure growth jumping from 2% to 9% and highest-exposure job share rising from 0% to 30%, supporting the claim that AI capabilities are advancing rapidly (cognizant.com).
Contradicting Evidence
Cognizant itself warns exposure does not equal inevitability
The report explicitly cautions that its scores represent a theoretical maximum based on AI capability alone and do not account for enterprise adoption rates, regulatory barriers, quality control requirements, or the pace of organizational change (cognizant.com).
Anthropic's real-world data shows augmentation dominates over automation
A January 2026 Anthropic study analyzing actual AI usage found a roughly 50/50 split between augmentation and automation tasks, with a slight edge toward augmentation — suggesting AI is enhancing human work more than replacing it, and that broader displacement unfolds over years, not months (axios.com).
IMF and OECD estimates are significantly lower
The IMF estimated in January 2024 that roughly 60% of jobs in advanced economies may be impacted by AI, well below Cognizant's 93%. The OECD's 2023 Employment Outlook placed 27% of employment in high-risk-of-automation occupations and noted limited evidence of large-scale displacement so far (imf.org, oecd.org).
OpenAI/UPenn research found narrower task-level exposure
A March 2023 study estimated about 80% of U.S. workers have at least 10% of their tasks exposed to AI and only 19% have 50% or more exposed — a materially lower threshold than Cognizant's broad 93% "impacted" framing (openai.com).
McKinsey projects full automation decades away
McKinsey's 2023–2024 analysis projects that 50% of current work activities could be automated sometime between 2030 and 2060, with a midpoint around 2045 — indicating that technical capability far outpaces real-world deployment timelines (mckinsey.com).
MIT research highlights economic feasibility constraints
A January 2024 MIT/CSAIL study found that only about 0.4% of U.S. wages tied to computer-vision tasks are currently economical to automate at today's costs, underscoring the substantial gap between what AI can theoretically perform and what is financially viable for businesses to implement (MIT CSAIL).
Report an Issue
Found something wrong with this article? Let us know and we'll look into it.