Experts tell House panel federal statistics must track AI at the task level

A witness testified that federal labor statistics are not currently designed to track how artificial intelligence is changing the nature of work.

Rivanna Shifuddin, a labor economist and research fellow at the Marketa Center at George Mason University, told the subcommittee that ‘‘AI transforms work differently than previous technologies’’ because it often changes specific tasks inside jobs rather than eliminating entire occupations. She said the Current Population Survey contains no questions about AI use or task changes and recommended an AI supplement to capture whether workers use AI, which tasks are affected, and whether AI is augmenting or automating those tasks.

Shifuddin also proposed linking farm-level adoption data to worker outcomes through existing Census programs and producing coordinated annual federal reports to give Congress a comprehensive evidence base. ‘‘The decisions this committee makes about workforce policy will shape whether AI delivers broadly shared prosperity. But good policy requires good data, and we currently don’t have it,’’ she said.

Committee members pressed for specifics on small-business adoption and tax incentives; Shifuddin suggested tax expensing provisions to equalize human-capital investment and encouraged targeted, low-cost data additions rather than expensive redesigns of federal surveys.

No formal agency commitments were announced during the hearing.