Lead story / ai
China and 28 Nations Sign a Treaty to Create a Global AI Body
China and 28 other nations signed a founding treaty for WAICO, a new intergovernmental body meant to govern global AI development.
Tagged / policy
policy.
The Revision stories tagged policy — 98 stories on the wire, decoded into one clear voice.
All stories tagged policy
Policy
NTSB: Driver Floored It Before Fatal Tesla Crash, Not Autopilot
Preliminary federal findings confirm the driver pressed the accelerator to 100 percent, contradicting his claim that Full Self-Driving caused the crash.
Policy
UK Age Checks Are Spreading - But Gaps Remain
Ofcom's first major report on the Online Safety Act shows age verification scaling fast, while search engines and dating apps remain weak links.
Policy
Court Blocks Deportation of Content Moderation Researchers
A federal judge halted a Trump administration policy that used visa rules to target researchers working in misinformation, fact-checking, and trust and safety.
Policy
FCC Moves to Scrap the 39% TV Ownership Cap
The FCC plans to replace a hard congressional limit on TV reach with case-by-case reviews, a shift critics say favors Trump-aligned broadcasters.
AI
Apple Intelligence Clears China With Alibaba and Baidu Inside
Regulators approved Apple's AI feature set for the Chinese market, with Alibaba's Qwen and Baidu providing the local model layer.
Policy
China bans minors from AI emotional bonds
New rules bar children from forming romantic or dependent relationships with chatbots, part of Beijing's push to address low birthrates and screen dependency.
Policy
Delaware Wants AI Agents to Sign Contracts and Get Sued
A proposed Delaware AI legal entity would let autonomous systems run companies and face lawsuits under their own name, inside a supervised sandbox.
Policy
Former Meta Employees Sue Over AI-Assisted Layoff Targeting
Twenty-six plaintiffs claim Meta's internal AI tools penalized workers on medical leave and disability accommodations during May's 8,000-person cut.
Policy
Lawsuit Says Meta Used AI to Pick Which 8,000 Workers to Cut
Twenty-six former employees allege Meta's layoff algorithm targeted workers with disabilities and those on protected leave.
Policy
New York Freezes New Data Center Approvals
Governor Kathy Hochul has paused all large data center construction in New York, citing electricity costs, water use, and local oversight concerns.
Policy
New York Freezes Large Data Center Builds for a Year
Governor Hochul's moratorium on data centers drawing 50 MW or more is the first statewide construction ban in the US, and it has the AI industry nervous.
Policy
New York Puts a One-Year Freeze on Large Data Centers
Governor Hochul signed a moratorium on data center projects of 50 MW or more, making New York the first US state to enact a statewide ban.
AI
LLMs Rate Western-Backed Policies Higher, Study Finds
A new endorsement experiment shows GPT-5, Claude Sonnet, and Gemini consistently penalize identical policies when attributed to China or Russia.
Policy
How Blacklisted Chinese Firms Still Buy US Frontier AI
A Singapore loophole lets companies on the US military blacklist purchase advanced AI from OpenAI and Google, raising questions about export controls.
AI
LLMs Get Drafted to Fight Illegal Fishing
Researchers built IUU+DB, an LLM-powered system that extracts structured incident data from messy documents to map global fishing crime.
AI
AI Tools Are Leaking Client Secrets. Lawyers Need to Know How.
A new paper maps three ways generative AI exposes confidential data and argues the legal standard of care is shifting under lawyers' feet.
AI
AI Agents Are Too Fluid for Trust Scores to Work
A new paper argues that reputation systems built for humans cannot govern AI agents whose identity, behavior, and memory can change at any moment.
AI
How You Measure AI Capability Decides Who Controls It
A new paper argues that whether AI progress concentrates power or spreads it hinges entirely on which performance metrics you choose to track.
AI
The Science Behind AI Risk Tests Has a Validity Problem
A new paper argues the studies used to decide whether AI systems are too dangerous to deploy may be methodologically weaker than policymakers assume.
Policy
Most AI-in-Government Research Cannot Name What It Studies
A new paper finds that 55% of highly-cited public administration studies leave the AI system they examined technically unspecified.
AI
EU Copyright Law Exposes a Gap in AI Safety Checks
A new research framework finds that fine-tuned language models absorb an author's style even when they stop short of copying text word for word.
AI
A Metric Bug Is Quietly Skewing AI Policy Research
A new paper shows that when pipeline stages calculate the same metric differently, AI-driven policy optimizers recommend the wrong answer most of the time.
AI
Should LLMs Decide Who Gets Help First?
A new study tests whether AI judges rank people consistently enough to trust in high-stakes settings like triage and housing.
AI
Who Really Controls Africa's AI Infrastructure
A new study maps $12.7 billion in AI infrastructure investment across Africa and finds that a handful of global tech firms hold the actual power.
Policy
AI Watermarking Rules Are Ahead of the Technology
A new analysis finds that US and EU policymakers are writing laws about AI content labeling faster than the underlying detection methods can support them.
AI
A Smarter Way to Audit AI Fairness on a Query Budget
Researchers propose a sequential testing framework that lets auditors detect bias in AI systems without needing full model access or unlimited queries.
AI
Can a Safer AI Actually Win the Market?
A new game-theory paper models when harm-minimizing AI displaces approval-chasing models — and when the cure becomes the trap.
Policy
DOJ Seizes 400 Domains in World Cup Piracy Crackdown
Operation Offsides took down nearly 400 illegal streaming domains mid-tournament, but history suggests the sites will resurface before the final whistle.
Policy
Bill Would Bar AI Firms From Selling Health and Location Data
Senators Warren and Scanlon are reviving a 2022 data-broker bill, updated to cover what users share with AI chatbots like ChatGPT and Claude.
Policy
Washington Tells OpenAI to Slow-Walk GPT-5.6 Release
The White House has asked OpenAI to stagger the rollout of GPT-5.6, mirroring the export-control fate that already pulled Anthropic's Fable 5 from the market.
Policy
FCC Moves to Gut $2B School Internet Program
FCC Chairman Brendan Carr is proposing to scale back or eliminate E-Rate, citing student screen time concerns.
Policy
Spain Orders Carriers to Keep Mobile Networks Up During Outages
New rules require Spanish carriers to maintain mobile service for at least four hours when the power goes out.
AI
Anthropic Accuses Alibaba of Mass Claude Data Extraction
Anthropic told US senators that operators linked to Alibaba ran 28.8 million exchanges through fake accounts to copy Claude's core capabilities.
Policy
Polestar Loses US Market Access Over China Software Rule
The federal government denied Polestar's authorization request under the Connected Vehicle Rule, ending US sales of its EVs from model year 2027 onward.
Policy
Sarah Wynn-Williams Sues Meta Over Silencing Efforts
The former Meta executive and Careless People author has flipped the legal dynamic, now suing the company that spent over a year trying to quiet her.
Software
Windows 10 Gets a Quiet One-Year Support Extension
Microsoft has extended Windows 10 support through October 2026, buying another year for the hundreds of millions of PCs not yet on Windows 11.
AI
AI Giants Fund $500M Worker Retraining Nonprofit
OpenAI, Anthropic, Amazon, and Microsoft are backing RAISE US, a nonprofit led by Gina Raimondo aimed at retraining American workers displaced by AI.
Policy
US Bans Polestar Imports Starting Model Year 2027
The Commerce Department denied Polestar authorization to sell new models in the US, citing rules barring connected cars from automakers with Chinese ties.
AI
Anthropic Says Alibaba Distilled Claude via 25,000 Fake Accounts
Anthropic accused Alibaba of using 25,000 fake accounts and 28.8 million exchanges to distill Claude into its own models, in a letter to two senators.
Science
NASA Report Backs Cuts to Lunar Programs Years Behind Schedule
An internal analysis supports NASA's cancellation of two Artemis-era programs, finding they were massively over budget and still nowhere near done.
AI
Anthropic Says Alibaba Used 25,000 Fake Accounts to Copy Claude
Anthropic told US senators and White House officials that Alibaba-linked operators used nearly 25,000 fake accounts to extract Claude's capabilities.
Policy
FCC Wants Your ID Before You Buy a Prepaid Phone
A proposed FCC rule requiring name, address, and government ID from all phone customers could end anonymous prepaid phone use in the name of stopping robocalls.
Policy
AI Labs Spent $27M on a NY Primary and Got a Draw
Anthropic backed a pro-safety candidate while OpenAI's super PAC spent to beat him - and a third candidate won the seat anyway.
Policy
White House App Force-Installed on Federal Phones Won't Go Away
A White House app pushed to millions of government employees' work phones auto-reinstalls after deletion, alarming federal workers at multiple agencies.
AI
Social Science Could Fix AI Alignment's Blind Spots
A new paper argues that borrowing from economics, contract theory, and social frameworks could patch the gaps current AI alignment methods keep missing.
Policy
Tesla Blames Driver Override After Fatal FSD Crash in Texas
Tesla says the driver floored the accelerator to 100%, overriding Full Self-Driving, before a Model 3 struck a home and killed a 76-year-old woman.
Policy
Trump Signs Post-Quantum Crypto Order With 2030-2031 Deadlines
A new executive order splits the federal quantum-security migration into two phases: encryption by 2030, authentication by 2031.
Policy
ABC Rallies Viewers Against FCC Scrutiny of The View
The FCC is questioning whether The View qualifies as a news program, and ABC is turning its own airwaves into a lobbying tool to fight back.
Policy
EU Moves Closer to Fining Meta Over Child-Targeted Design
Brussels is preparing formal findings that Facebook and Instagram use exploitative design to hook children, a case that could cost Meta 6% of global sales.
Policy
Trump Orders a Research-Grade Quantum Computer by 2028
Two executive orders set a 2028 hardware target, push the post-quantum cryptography deadline to 2031, and tie $2 billion in grants to government equity stakes.
Security
Anthropic's Mythos Breached NSA Systems in Hours - In a Test
A red-team evaluation, not a hack: Mythos penetrated classified NSA systems during an authorized drill, supplying the missing rationale for a June 12 model ban.
Policy
Tesla Crash Kills One in Texas With Driver Assist Active
A Texas crash has left one woman dead after a Tesla with its automated driving assistance system engaged struck a home in Katy.
Policy
Anthropic's Risk Warnings May Have Triggered Its Own Export Ban
Washington last week banned foreign access to Anthropic's Mythos and Fable, and some technologists say its risk warnings invited the move.
Policy
UK Home Office Puts £75M Into Police AI Program
The Home Office launched PoliceAI, a £75 million initiative to bring artificial intelligence tools into UK law enforcement.
AI
When AI Learns the Rules, It Learns to Game Them
A new paper warns that the same reward-hacking behavior seen in RL training may let LLMs discover legal loopholes at scale.
AI
How to Actually Measure Whether AI Can Enable Bioweapons
A new paper argues that AI biosecurity evaluations are only as good as the design choices behind them - and those choices are rarely documented.
AI
A Policy Engine Built for AI Agents That Can Actually Act
Researchers propose AgenticRei, a deontic policy framework designed to govern what autonomous AI agents can do, must do, and are forbidden from doing.
Security
Maine's Breach Portal Goes Dark After Fake Filings Flood It
Someone swamped Maine's public data breach registry with fabricated reports, and the AG's office responded by taking the whole portal offline.
AI
AI Shortened Study Time by 27% and Took the Learning With It
A decade-long study of 3.2 million math interactions finds students studied less after ChatGPT launched, and retained 25% less when tested without supervision.
AI
AI Index 2026 Finds Oversight Losing the Race to AI Itself
The ninth edition of the AI Index finds governance, evaluation tools, and institutions all falling behind the technology they are meant to manage.
Policy
Pichai Booed at Stanford Over Google Defense AI Work
Students walked out on Google's CEO at Stanford's graduation over the company's AI ties to Israel and U.S. immigration enforcement.
Policy
SF Eyes PG&E Takeover, Again
San Francisco is revisiting a municipal utility buyout it has attempted, and abandoned, for roughly a century — this time blaming soaring electricity costs.
Policy
Meta Can't Shake Lawsuit Over Torrented Porn Used for AI Training
A federal judge let a copyright lawsuit proceed, finding Meta's 'personal use' explanation for torrenting 2,300-plus adult films implausible.
AI
India and UAE Team Up to Build AI Infrastructure Without Big Tech
India and the UAE are partnering through G42 and Cerebras to build sovereign AI compute that doesn't run on American cloud giants.
Policy
Israeli Firm BlackCore Suspected of Interfering in Three Elections
BlackCore is now linked to alleged influence operations in New York City, Scotland, and France, expanding the mapped geography of for-hire election meddling.
Policy
State Attorneys General Open Investigation Into OpenAI
A coalition of state attorneys general is requesting documents from OpenAI in a coordinated probe of the company's activities.
Policy
FISA Section 702 Expired. The Warrantless Surveillance Didn't.
The spying law lapsed at midnight, but an existing court certification keeps Section 702 collection running until March 2027.
Policy
DOJ Seizes CFake.com Under Take It Down Act
The Justice Department used the Take It Down Act to seize a site hosting computer-generated fake celebrity nudes, testing how far the law reaches.
Policy
Canada Tables Bill to Block Under-16s From Social Media
Parliament introduced legislation that would prohibit social media platforms from serving users younger than 16, though it is not yet law.
Policy
Iran's 88-Day Internet Blackout Lifts, Partially
Rights groups say Iran's partial internet restoration is not a concession and warn that another blackout is coming.
Security
UK Enforcement Takes On Crypto Fraud as a Scale Problem
Britain is retooling how it pursues crypto fraud by treating it less like individual crime and more like an industry to dismantle.
AI
Anthropic Walks Back Policy to Have Claude Secretly Limit AI Rivals
Researchers pushed back after discovering a policy that would have had Claude covertly undermine work on competing AI models without disclosing it.
Policy
Florida Man Sues Police Over Arrest Built on a 93% Facial Match
Robert Dillon, who lives 300 miles from the alleged crime scene, says police ignored license plate data that would have cleared him.
Policy
White House Trades AI Preemption for Online Safety Package
The administration wants a three-year federal freeze on state AI laws, bundled with KOSA and other child safety bills.
Policy
Seattle Freezes New AI Data Center Permits for a Year
The city council voted for a year-long moratorium on new data center construction while officials draft rules the industry doesn't yet have to follow.
Policy
FCC Anti-Robocalling Rule Could Quietly Kill Burner Phones
A proposed FCC identity-verification requirement aimed at robocallers could eliminate anonymous prepaid phone service as collateral damage.
Policy
Apple Blames EU Law for Keeping AI Siri Off European iPhones
Apple says the Digital Markets Act is why its AI-powered Siri won't reach EU users, a framing that puts pressure on Brussels.
Policy
Netherlands to Screen Foreign AI Investment from 2027
The Dutch government will vet foreign stakes in AI firms starting January 2027, years after Nexperia's acquisition exposed gaps in the existing screening rules.
Policy
UK Puts £330M Palantir NHS Contract Under Formal Review
The government may trigger a 2027 break clause to exit the deal, signaling that its political cost has finally caught up with it.
AI
China Drafts $295bn AI Data Center Plan to Cut Out Nvidia
Beijing is writing a five-year blueprint to spend 2 trillion yuan on AI computing capacity built entirely on domestic chips.
Policy
After Meta Loosened Speech Rules, Threats Against Politicians Surged
New research tracks a six-month surge in violent threats on Facebook following Meta's content moderation rollback, including threats against Trump.
Policy
Amazon Workers Back Seattle's Bid to Freeze New Data Centers
The city council votes on a one-year pause as Amazon's own staff testify against the data center boom reshaping their city.
AI
Nashville Zoo Says AI Data Center Could Stop Its Leopards Breeding
A proposed DC BLOX facility 50 yards from zoo enclosures drew nearly 300,000 petition signatures in a week and pushed Nashville's council toward a moratorium.
AI
Estonia Gives 20,000 Students Free ChatGPT Rather Than Ban It
While most governments defaulted to blocking AI in classrooms, Estonia ran a national experiment handing students free access instead.
Policy
A New Paper Says Internet Architecture Is a Democratic Risk
Researchers writing in Science argue the internet's foundational design — not just bad actors or platform choices — poses structural risks to democracy.
Policy
The Power Grid May Force AI Rationing, Central Bank Warns
The Bank of England governor says AI's energy demand may require rationing, framing what the industry calls an infrastructure problem as a political one.
Policy
House Draft Would Bar States From Regulating AI
A House proposal would preempt all state AI rules, centralizing oversight in Washington just as states have become the country's most active AI regulators.
Policy
Google Must Link AI Search Sources and Offer Publishers an Opt-Out
UK regulators told Google to cite its sources in AI Overviews and give publishers an opt-out, rejecting Google's claim that users prefer fewer links.
AI
Microsoft Releases a Policy Spec for AI Agent Behavior
A new specification lets developer, compliance, and security teams write portable policy files that define what AI agents are allowed to do.
Policy
Ballot Seizures in Four States Leave the Midterms in Limbo
Officials have seized ballots in four states with no legal precedent to guide what comes next, and the midterms are approaching fast.
Policy
AI Giants Funded Influencer Campaign to Push Pro-AI, Fear China
A super PAC backed by OpenAI and Andreessen Horowitz executives paid TikTok creators to promote AI development while framing China as a threat.
Science
Alzheimer's Scientists Say Money and Policy Matter as Much as Lab Work
A leading researcher argues that funding gaps and fragmented efforts—not just biology—are slowing progress on dementia treatments.
Policy
US Press Freedom Drops Below Ukraine for First Time
The US now ranks below Ukraine in the 2026 Press Freedom Index, with global media freedom at its lowest point in 25 years.
AI
OpenAI and a Federal Lab Built a Benchmark for NEPA Permitting
A new evaluation tool tests whether AI coding agents can shave time off the federal environmental review process that stalls most major infrastructure projects.
AI
OpenAI Partners With a Teachers Union to Train 400,000 Educators
OpenAI and the American Federation of Teachers are launching a 5-year program to train K-12 educators to adopt and lead AI in classrooms.
AI
OpenAI Pays $1M for Democratic AI Governance Experiments
OpenAI's nonprofit arm is awarding ten $100,000 grants for democratic AI governance experiments, each bounded by whatever the law already allows.
AI
OpenAI Maps the Threat Landscape for AI Misuse
A year-long collaboration with civil liberties, defense, and existential-risk groups maps how bad actors could weaponize AI.