US bill proposes $5K tax credit per worker in AI training

A bipartisan bill introduced in the US Congress proposes a $5,000 tax credit per worker trained in artificial intelligence. The *AI Workforce Empowerment Act*, sponsored by Senators Mark Warner (D-VA) and Todd Young (R-IN), is the first American fiscal instrument to explicitly tie tax incentives to AI qualification.

What the bill proposes in detail

The mechanism is direct: a company that covers AI training costs for employees — certified courses, bootcamps, Department of Labor-recognized certification programs — can deduct up to $5,000 per trained worker from federal corporate income tax.

Retention clause: the worker must remain employed by the same company for at least 12 months after completing training. If laid off before that period (except for cause), the tax credit is clawed back proportionally.

Eligible courses: the DoL will publish a quarterly list of eligible courses and certifications. Provisional criteria include: minimum 40 hours of instruction, competency assessment at the end, and focus on practical applications.

Why the design matters as much as the amount

Skills policies classically fail for one reason: the company gets the incentive, runs a low-quality training, and fires shortly after. The *AI Workforce Empowerment Act* addresses this with two mechanisms: the retention clause and a curated course list. Per the Tax Policy Center, this dual mechanism is "the most sophisticated ever proposed for a professional qualification credit" since the Workforce Investment Act of 1998.

Sources

• Congress.gov — full text of the *AI Workforce Empowerment Act*
• Tax Policy Center — credit mechanism analysis (taxpolicycenter.org, Jun. 2026)
• McKinsey Global Institute — *The Future of Work in America* (Jan. 2026)
• Harvard Business School — *AI Productivity Gains in Knowledge Work* (2025)
• Bloomberg Government — legislative prospects analysis (bgov.com)