Democrats preparing to take control of Congress in January said they will push in 2027 for new AI rules targeting data-center electricity costs, independent testing of leading models, and corporate liability for AI harms. Senator Chuck Schumer, expected to become Senate majority leader, named protecting consumers from electricity-bill increases tied to data-center growth as a priority for next year. Representative Lori Trahan said she would seek independent expert reviews of widely used AI models and clearer rules on companies' responsibility for harm their products cause.
The agenda follows midterm results that turned voter unease about AI into a political force. NBC News reported that roughly 56 percent of voters said concerns about AI were crucial or important to their vote, a figure the Tech Oversight Project's Sacha Haworth said she would not have believed six months ago. Opposition to data centers, driven by worries over electricity, water and utility bills, ran through dozens of local ballot measures, almost all of which passed.
What Democrats cannot yet agree on is where to aim first. Incoming San Francisco representative Scott Wiener, who wrote California's state AI law, said "every Democrat in the House I'm speaking to wants to see AI regulation," while acknowledging "there may be different ideas about what that should be." Those ideas range widely: Senator Bernie Sanders has proposed banning artificial superintelligence outright, Senator Sherrod Brown has concentrated on data centers, and re-elected Senator Jon Ossoff has stressed the technology's catastrophic risks. Samuel Hammond of the right-leaning Foundation for American Innovation predicted Democrats "are going to have a hard time wrangling their own perspectives," though he also argued AI is "not nearly as left-right polarized as most other issues."
The party is also weighing how aggressive to be. Democratic leaders are reviewing a 181-page investigation checklist prepared by the Groundwork Collaborative that covers technology, AI, natural gas, health care, private prisons and elder care, and proposes hearings at which executives would explain consumer fees and business models. In October, Representative Zoe Lofgren and other Democrats asked OpenAI, Anthropic, Meta, Alphabet and SpaceX to disclose safety measures and identify cases where their tools acted outside intended design. Some moderates favor narrower investigations, wary of appearing anti-business.
The plans run directly against President Trump, who reiterated his support for "superintelligence" on November 6 and has repeatedly opposed mandatory AI rules, calling data centers an economic "golden goose." Committee control would let Democrats advance bills and compel testimony, but overriding a presidential veto requires two-thirds of each chamber, which they do not have. One existing measure may survive the shift: the bipartisan FRONTIER Act, co-sponsored by Representatives Jay Obernolte and Trahan, would mandate independent testing and published safety plans for the most advanced systems, though Obernolte's own re-election remained undecided as of the weekend. Alex Bores, the New York assembly member behind the RAISE Act, said pro-AI money had proved "toxic" to some candidates this cycle, noting losses for contenders backed by the super PAC Leading the Future.