AI backlash meets weak rules
Public sentiment on AI has flipped: polls show a majority of Americans now expect AI to harm them or shrink job opportunities, fueling political pressure on regulators. Lawmakers and regulators are experimenting with 'AI sandboxes' to let models be tested safely, but experts warn national policy frameworks lack clear accountability about who oversees the overseers — meaning the country that manages the backlash best may gain a strategic edge. (straitstimes.com) (forbes.com) (brookings.edu) (newsweek.com)
Quinnipiac surveyed 1,397 U.S. adults from March 19–23 with a ±3.3 percentage‑point margin of error and reported that 51% say they use AI for research while just 21% say they trust AI‑generated information most or almost all of the time. (poll.qu.edu) The White House published a National Policy Framework for Artificial Intelligence on March 20, 2026 that calls on Congress to enact AI legislation and recommends preempting some state AI laws. (natlawreview.com) Brookings contributors Tom Wheeler and Bill Baer wrote on March 31 that the White House framework “sidesteps” the central question of who oversees the overseers, arguing the document lacks clear institutional accountability for AI governance. (brookings.edu) The EU’s implementation guidance for the AI Act requires member states to establish or participate in at least one AI regulatory sandbox, with those sandboxes expected to be operational by August 2, 2026. (europarl.europa.eu) The UK’s Medicines and Healthcare products Regulatory Agency launched the AI Airlock sandbox for AI as a Medical Device in 2024 and has secured a £1 million investment to expand cohorts and produce case studies including Philips and AutoMedica. (assets.publishing.service.gov.uk) In the U.S., MITRE has developed a Federal AI Sandbox to train foundation models for agency use and the GSA launched a cloud‑based USAi platform to let federal agencies test AI tools at no cost. (mitre.org) A March 31 Newsweek opinion by Maha Hosain Aziz argues the country that combines rapid AI deployment with robust risk‑management and clear oversight will gain a geopolitical and economic advantage in the global AI race. (newsweek.com) Bloomberg reported that Amazon, Meta, Google and Microsoft plan a combined roughly $650 billion in AI infrastructure spending this year, while Quinnipiac found 74% of Americans say the government is not doing enough to regulate AI—together increasing political pressure on regulators. (bloomberg.com)
