News-May-05-2026

May 5, 2026 — Daily digest of AI law developments.

New Mexico Seeks $3.7 Billion from Meta as Phase 2 Trial Opens

The second phase of New Mexico's landmark child safety case against Meta Platforms began in Santa Fe on May 4, 2026, with the state seeking $3.7 billion in abatement costs and sweeping injunctive relief. The bench trial, expected to last three weeks, will determine whether Meta's actions constitute a public nuisance warranting product changes to Facebook and Instagram. New Mexico Attorney General Raúl Torrez had already prevailed in the first phase in March 2026, when a jury found Meta willfully violated the state's unfair practices act and awarded $375 million. The judge warned prosecutors against "overreach" in the remedies sought, which include extensive changes to how the company delivers its services in New Mexico.^[1]^[2]

See also: Full article | Case page

Ball & Buchanan Call for Bipartisan AI Security Action

Two former White House AI advisers from opposing administrations co-authored a New York Times op-ed on May 4, 2026, calling for bipartisan action on AI security risks. Dean Ball, a former Trump White House AI policy adviser, and Ben Buchanan, who served as President Biden's White House adviser for AI, urged Congress and the executive branch to work together on catastrophic AI risk mitigation. The op-ed recommends tighter export controls on advanced AI chips and mandatory safety audits for frontier AI models. The bipartisan appeal comes as the Trump administration is reportedly considering an executive order to create an AI working group for model vetting.^[3]

See also: Full article

Google DeepMind UK Staff Unionize Amid Military AI Concerns

More than 1,000 staff at Google DeepMind's UK-based operations voted to unionize with the Communication Workers Union (CWU) and Unite in April 2026, according to a letter made public May 4. The unionization comes amid ongoing concerns over DeepMind's involvement in a U.S. Department of Defense classified AI contract. Staff cited concerns about the ethical implications of military AI applications and the lack of worker input in decisions about DOD partnerships. The CWU and Unite will represent DeepMind workers in collective bargaining over working conditions, project assignments, and ethical review processes.^[4]

Google, Microsoft, xAI Join CAISI Early Model Access Program

Google DeepMind, Microsoft, and xAI agreed on May 5, 2026 to give the U.S. Commerce Department's Center for AI Standards and Innovation (CAISI) early access to their frontier AI models for pre-deployment security and capability evaluations. The three companies join OpenAI and Anthropic, bringing the total number of participating frontier AI labs to five. The voluntary cooperative model represents an alternative to mandatory regulatory frameworks like the EU AI Act. The announcement follows a bipartisan op-ed by former Trump and Biden AI advisers calling for mandatory safety audits and tighter export controls.^[5]