Data Center Backlash
7 in 10 Americans Don’t Want a Data Center Near Them. A Gallup poll finds overwhelming opposition to AI data center construction across the political spectrum, with half of opponents citing excessive resource consumption.
Anti-Data Center Activists Are Taking on Big Tech and Winning. More than 70 data center projects were rejected or restricted in the first four months of 2026 alone, exceeding the total for all of 2025. Broad coalitions now span 188 groups across 17 states.
States Are Trying to Block Cities from Regulating AI Data Centers. An increasing number of state legislatures are pursuing preemption laws that would strip local governments of the power to restrict data center development, even as residents organize bans and moratoriums.
Utah Residents Protest a 9-Gigawatt AI Data Center Backed by Kevin O’Leary. A planned 40,000-acre campus near the shrinking Great Salt Lake has become a flashpoint for the tension between AI infrastructure ambitions and local environmental concerns, including water use and migratory bird habitat.
DOJ Weighs Intervening in NAACP’s Clean Air Act Lawsuit Against xAI. The Justice Department has until June 15 to decide whether to join a federal lawsuit alleging xAI is running dozens of unpermitted gas turbines to power its Colossus 2 data center in Mississippi, disproportionately affecting nearby Black neighborhoods.
Meta Data Center Allegedly Muddies Georgia Town’s Drinking Water. The EPA has launched an investigation after a congresswoman brought jars of dirty water to a hearing, alleging a Meta data center project contaminated the local water supply.
EU Environment Agency Chief: Don’t Let Big Tech Hide the Ecological Cost of AI. The head of the European Environment Agency is urging the EU to mandate disclosure of data center energy consumption and emissions as part of a sustainability label alongside existing safety and transparency rules.
Deepfakes & the Take It Down Act
DOJ Announces First Arrests Under the Take It Down Act for AI Deepfake Pornography. Two men were charged in separate cases for allegedly creating and distributing AI-generated non-consensual intimate images, marking the first criminal enforcement under the new federal law.
FTC Begins Enforcing the Take It Down Act. The FTC explains the new platform compliance obligations, which require removal of non-consensual intimate images (including AI-generated ones) within 48 hours of notification, and launches a reporting portal at TakeItDown.ftc.gov.
What’s Next for the Nation’s First AI Deepfakes Law. Axios examines the enforcement and implementation challenges ahead, including the transparency gap around how platforms will decide what gets removed and what stays up.
Google Makes It Easy to Deepfake Yourself. Google Flow’s new avatar and selfie-video feature normalizes consumer deepfake creation, blurring the line between consent-based digital doubles and the broader ecosystem of likeness misuse.
Policy & Regulation
European Commission Opens Consultation on High-Risk AI Classification Under the EU AI Act. The Commission published draft guidelines to help providers and deployers determine whether their AI systems qualify as “high-risk,” with practical examples across education, employment, biometrics, and law enforcement. Public comments are open until June 23.
Pope Leo XIV Calls for Strict AI Regulation in First Encyclical. The 42,000-word document warns about AI’s potential to dehumanize society, calls for ethical constraints on AI in work and warfare, and cautions against the concentration of digital power in major tech firms.
Colorado Scraps Its AI Act, Replaces It with a Narrower Automated Decision-Making Law. Colorado’s new law, effective January 2027, shifts from broad AI regulation to targeted rules on automated decision-making in consequential contexts. Business groups counted it as a major legislative win.
Lawfare: Should Congress Have to Approve Major Government AI Deployments?. Cullen O’Keefe argues that advanced AI systems used in abuse-prone government functions like surveillance and public administration should require congressional authorization, drawing on democratic accountability principles.
The U.S. Has 1,200 AI Bills and No Good Test for Any of Them. A policy analysis of America’s fragmented regulatory landscape, noting that state-level rules risk pushing companies to friendlier jurisdictions while open-source alternatives create built-in evasion routes for any single jurisdiction’s restrictions.
Ethics & Safety
Peer-Reviewed Study: Mass Media Reports of Psychiatric Harms from AI Chatbots Center on Youth Suicide. A scoping review in JMIR Mental Health finds that reporting on AI-related psychiatric harms is concentrated around severe outcomes, particularly suicide deaths among young people, and is frequently framed around regulatory and corporate accountability failures.
Brookings: Treat Unsafe AI Companion Bots Like Defective Products, Not Speech. This report argues for a recall-style regulatory regime for AI companion bots, drawing on public health frameworks to address addictive design, erosion of social skills in children, and absent safety guardrails.
The Blind Spot in AI Safety. A critique of how current AI safety research frames frontier-model failure, arguing that governance debates may be over-indexing on certain failure modes while missing institutional and real-world deployment risks.
METR Frontier Risk Report: February to March 2026. The independent evaluator reports on agentic behavior, rogue-deployment concerns, and monitoring challenges in frontier models, linking technical evaluation findings directly to governance-relevant risk scenarios.
Calling AI “Intelligent” Creates Accountability Gaps. Brookings warns that human-like language around AI systems shifts perceived responsibility away from developers and institutions, weakening policy oversight and ethical accountability.
Economics & Employment
Meta Lays Off 8,000 Workers as Zuckerberg Warns “Success Isn’t a Given” in the AI Race. The 10% workforce cut is directly tied to AI-focused restructuring, making it one of the clearest examples of how AI investment is reshaping headcount at major tech firms.
Dallas Fed: AI Helps Experienced Workers but Squeezes Entry-Level Ones. Federal Reserve research finds that employment for 22-to-25-year-olds in AI-exposed jobs fell 6% since late 2022, while employment for workers over 30 in the same roles grew 6-13%. The traditional model of learning on the job may be breaking down.
AI-Exposed Industries See Output Gains, but Most Benefits Flow to Capital, Not Workers. Research using U.S. administrative data shows AI exposure boosts output by 10% and employment by 3.9%, but only 29% of the gains go to labor. The rest accrues to capital and profits.
Indeed’s Chief Economist: AI-Exposed Sectors Are Actually Seeing Job Growth. A more optimistic read on the labor market, suggesting that sectors most exposed to AI are also seeing rising demand for AI-fluent workers, complicating the straightforward displacement narrative.
Research & Copyright
Legal Analysis: Elsevier v. Meta and the “Pirated Books” AI Training Lawsuit. Holland & Knight breaks down the class-action lawsuit alleging Meta torrented pirated books from shadow libraries to train Llama. The case could reshape the fair use defense for AI developers.
Sony Music Seeks to Add 30,000 Tracks to Its Copyright Lawsuit Against Udio. After gaining access to Udio’s training data during discovery, Sony identified tens of thousands of additional copyrighted recordings it wants added to the complaint.
New Paper: How Generative AI Is Reorganizing Labor Demand. Using a nationwide U.S. job-postings dataset, this research moves beyond “jobs lost vs. gained” to examine how firms are restructuring task architecture and organizational roles as generative AI spreads.
Whose Voice Counts? Mapping Public Comments on the AI Action Plan. An analysis of 10,000+ public letters reveals that individuals prioritize job loss, energy costs, and artist rights, while the private sector focuses on innovation and security. The final plan under-represents public concerns about inequality.
Last Updated: 2026-05-26 07:24 (California Time)