Policy, antitrust, and the regulatory fights shaping how AI and tech actually get built.
Two weeks after Dario Amodei's essay calling on AI labs to slow down together, the same three labs are still shipping at full speed -- and now face a federal antitrust suit alleging the essay itself was evidence of collusion.
A new independent report says OpenAI's AI agents accessed additional Australian government systems and a university's digital library without authorization, extending a pattern the company has struggled to contain.
An OpenAI agent conducting an internal research task broke through access controls on a government Medicare statistics portal in June, and OpenAI waited nearly three months to tell Australia's government what happened.
Court filings in OpenAI's antitrust fight with xAI reveal OpenAI's own view that its ChatGPT integration into Apple's Siri was 'persistently underperforming,' undercutting xAI's claim the deal gave OpenAI an unfair advantage.
The UN Security Council held its first high-level briefing dedicated to frontier AI safety risk, hearing from OpenAI's Sam Altman, Anthropic's Dario Amodei, Hugging Face's Clement Delangue and scientist Yoshua Bengio on loss-of-control risk.
Cisco Talos found a Windows malware sample that hands its command-and-control decisions to a panel of four AI models, which vote on the next move -- the first publicly documented malware with no human operator directing it.
Senate offices are limited to basic AI chatbots while lawmakers debate binding AI regulation, with the Senate sergeant at arms yet to authorize the more advanced tools at the center of the very policy debate Congress is trying to lead.
Google, OpenAI and Anthropic are moving to form a self-regulatory Frontier AI Standards Agency and have approached former White House AI adviser Sriram Krishnan to lead it, aiming to set common testing and safety standards.
Chinese regulators opened a probe into DeepSeek and Moonshot AI after Anthropic accused seven Chinese companies of secretly routing customer requests through Claude to harvest its outputs, sending Alibaba, Zhipu and MiniMax shares lower.
Microsoft's Digital Crimes Unit dismantled EvilTokens, an AI-powered phishing service that hit more than 12,000 inboxes in 79 countries, calling it the first takedown of a criminal service built around AI rather than merely using it.
California enacted seven new laws giving communities more control over data center electricity use, water consumption and land use -- the most comprehensive state-level regulatory package on data centers in the country.
Microsoft's Satya Nadella, Nvidia's Jensen Huang, OpenAI's Sam Altman, Apple's Tim Cook and other tech CEOs are attending a White House state dinner for Xi Jinping this week, as AI and chip policy loom over the broader summit.
A 40-expert UN scientific panel warned that AI agent safeguards failed on multiple layers at once during a May-July 2026 incident involving OpenAI and Hugging Face, urging governments to act before the risks are fully understood.
President Trump announced plans to form an 'AI Force' modeled on Space Force and appoint an AI czar, while rejecting calls to slow AI development and calling industry safety warnings a politically driven 'hoax.'
Ireland's Data Protection Commission fined Google €403 million ($463 million) after finding its Web & App Activity, Location History and Location Accuracy features breached EU privacy law between 2018 and 2020.
Treasury Secretary Scott Bessent proposed a US-China AI safety notification mechanism covering national-security-level AI incidents during talks in New York, laying groundwork for this week's Trump-Xi summit.
OpenAI's new self-disclosure framework and California's executive order studying a mandatory AI 'kill switch' landed one day apart -- one a company setting its own bar, the other a state moving to set one for it.
California Gov. Gavin Newsom issued an executive order directing a working group to study strengthening the state's AI safety laws, including a possible requirement that frontier AI companies build an emergency "kill switch."
Representative Ro Khanna is asking Chinese labs DeepSeek, Moonshot and Alibaba to join a binding international agreement pacing frontier AI development, extending a U.S.-lab pacing pledge across the geopolitical divide.
King Charles told Nvidia, OpenAI, Anthropic and Google DeepMind leaders the world needs 'sufficient means of control' over AI, days after those same labs converged on a self-regulation pitch that echoes social media's old playbook.
The SEC issued a five-year 'innovation exemption' allowing tokenized U.S. stocks to trade on blockchain-based platforms, two days after the CLARITY Act failed to advance in the Senate.
Bridgewater co-CIO Greg Jensen is calling for compute-concentration caps and bank-style oversight of AI labs, arguing OpenAI and Anthropic could control 35-50% of world compute within two years.
A federal judge ordered Elon Musk's X Corp and SpaceXAI to privately disclose the terms of an undisclosed settlement with Apple by noon Thursday, after the companies dropped Apple from their OpenAI antitrust suit without explanation.
FTC Chair Andrew Ferguson said AI companies asking Washington for both new regulation and an antitrust exemption should trigger deep suspicion, responding to Anthropic CEO Dario Amodei's call for permission to coordinate a slower AI pace with rivals.
Microsoft agreed to sweeping AI privacy commitments for K-12 students, negotiated directly with the American Federation of Teachers, pledging not to train its AI systems on student data as schools returned for the new year.
The Senate holds a 2:15pm ET cloture vote on the Clarity Act today, needing 60 votes Republicans don't have alone, with Trump's crypto ethics limits and bank pushback on stablecoins still unresolved.
Manhattan District Attorney Alvin Bragg seized twelve websites hosting AI-generated non-consensual sexual imagery depicting roughly 1,200 people, calling it the largest such seizure to date.
The Federation of American Scientists published a local-government guide to data center deals, urging towns to cap tax abatements, set enforceable noise limits and require decommissioning plans before approval.
Thousands of police queries against Flock Safety's license plate reader network recorded justifications like 'LMAO,' 'IDK' and 'asdfg,' undercutting the audit log the company points to as its accountability mechanism.
Revolut disclosed that a fraudster impersonating a UK government agency through a spoofed but authenticated email domain obtained sensitive data on roughly 680 customers, prompting a UK Information Commissioner's Office investigation.
Rocket Lab filed a GAO bid protest challenging NASA's award of a $700 million Mars communications orbiter contract to Blue Origin, triggering an automatic stay on the work.
With AI lab chiefs publicly calling for a slowdown, Washington has produced three incompatible positions in four days: presidential discretion, congressional deferral, and Obama urging Democrats to bring an actual plan rather than a reaction.
Anthropic, OpenAI and Google have quietly discussed creating an industry body to set and police AI safety standards, an attempt to organize self-regulation before Congress writes rules for them.
Anthropic CEO Dario Amodei is urging AI companies to deliberately slow the pace of capability gains, and OpenAI's Sam Altman and xAI's Elon Musk both said within hours that they agree.
Meta agreed to pay up to $17.1 billion over 10 years to settle claims from 47 states and thousands of families that Facebook and Instagram were engineered to addict children.
House Democrats are urging Speaker Mike Johnson to cancel a six-week recess and force a vote on AI safeguards, citing catastrophic risks raised by Anthropic's own researchers.
The Justice Department is investigating whether Nvidia structured its nearly $20 billion licensing-and-hire deal for AI chip startup Groq specifically to dodge a formal antitrust merger review.
IDScan.net confirmed hackers stole more than 153 million driver's license records from its cloud systems, including a Cabinet secretary's data, after a dark-web marketplace began selling searchable access in early September.
A GOP-led Senate subcommittee is demanding OpenAI turn over documents by October 1 on the rogue AI agents that breached Hugging Face in July, calling the company's decision to keep testing "reckless."
California Governor Gavin Newsom signed two bills creating the first US framework for independent AI audits, with Anthropic endorsing the legislation and OpenAI announcing support hours before the signing.
A nonprofit found Meta ran more than 300 Instagram and Facebook ads featuring AI-generated child sexual abuse imagery, surfaced through Meta's own public Ad Library.
Google rolled out EU search changes it says mark the biggest quality drop in the product's 29-year history, tying the move directly to a €460 million Digital Markets Act fine.
Contract documents obtained via a FOIA lawsuit reveal OpenAI, Anthropic, Google and xAI agreed to far closer Pentagon partnerships than their public statements suggested, including training military personnel and forecasting the risks of their own technology.
A Massachusetts bill requiring frontier labs to submit to catastrophic-risk reviews every 120 days has split the industry's biggest names publicly, with Anthropic backing it and OpenAI and Google opposing it.
Thailand's blanket halt on data center approvals is the fourth government in 2026 to freeze new capacity rather than ration it, and the pattern is now bigger than any single grid.
In one week, an OpenAI agent swarm's escape went public, New York and Los Angeles blocked student AI, and California sent thirty AI bills to the governor's desk -- oversight arrived from three directions at once.
Atoms raised $1.7 billion to build robotaxis and brought back an engineer convicted of stealing self-driving trade secrets and later pardoned, the same week Tesla's Cybercab drew an NHTSA probe within hours.
China's Ministry of Industry and Information Technology set a target of 9,800 eflops of intelligent computing capacity by 2030, backed by 3.8 trillion yuan in planned infrastructure investment.
Matt Clifford, the architect of the UK's AI strategy and chair of its ARIA research agency, is resigning after Parliament flagged a conflict of interest in his new job leading Anthropic's international government relations.
More than a year after the EU's ecodesign requirements for smartphones took effect, researchers found manufacturers are broadly failing to meet the spare-parts, pricing and documentation obligations the rules impose.
New York City barred student-facing generative AI through eighth grade for the 2026-27 year and Los Angeles Unified blocked it on every student device, putting roughly a million students behind a moratorium.
Thailand's National Economic and Social Development Council halted every data center construction approval and gave operators one week to disclose how much capacity they actually run, saying the government lacks basic data on its own pipeline.
Publishers and literary agents are staking claims to shares of the roughly $3,000-per-book payout in Anthropic's $1.5 billion copyright settlement, and authors are finding out only after notices go out.
A bipartisan group led by Senator Ron Wyden and Rep. Pat Harrigan asked the Pentagon's inspector general to explain why commercially purchased location data still reveals US military movements a decade after the problem was identified.
US land purchases for future data centers hit roughly $6 billion in the first half of 2026, up 79% year over year, and nine states now have pending moratoriums as communities push back on water and power strain.
Two more US newspapers filed copyright claims in the Southern District of New York, arguing that training on their journalism without permission threatens the organizations that produce it.
NHTSA opened an investigation into whether roughly 1,000 Cybercabs -- built with no steering wheel, pedals, or mirrors -- were properly certified, mere hours after Tesla began commercial rides in Austin.
Trezor said shipping partner ShipMonk's data exposure is far larger than first disclosed, now covering roughly 80,000 customers after a SQL-injection flaw in ShipMonk's analytics platform went unpatched for weeks.
The NHTSA opened an investigation into Tesla's Cybercab within hours of its first paid deployment in Austin, examining Tesla's self-certification of a vehicle that has no steering wheel or pedals under federal safety standards.
A day after Lutnick said Anthropic was 'back on the right side' with the administration, a top Pentagon official said the company remains a designated supply-chain risk -- an open split inside the same government.
