OpenAI removes live internet from frontier training
OpenAI says its frontier-model training environments will no longer use the live internet after security testing exposed containment weaknesses.
Policy — Category
US Orders DeepSeek Removed From Federal Civilian Systems
A DHS order bars civilian agencies from buying DeepSeek technology and requires removal of third-party products built on its implementations.
Anthropic Commits 5% of Shares to Public Benefit Fund
Anthropic will place 5% of its equity in a wholly owned fund, tying future public-benefit spending to the company’s growth and governance.
Anthropic Asks Law Firms to Shape Agent Negotiation Rules
Anthropic is consulting legal experts on how AI agents should negotiate prices, complaints and contracts before wider commercial deployment.
Meta Terms Could Permit Training on Linked Business Sites
Meta’s upcoming commercial terms may let it crawl and use linked business websites for AI development, with no clear opt-out.
Google and Singapore Build On-Device Scam Detection Model
Google DeepMind and Singapore’s GovTech have built an on-device Gemma-based system for localized scam conversations; its rollout within messaging platforms and applications will be phased and consent-based.
Nadella Calls for an Emergency Brake on AI Models
Microsoft CEO Satya Nadella urges firms to treat advanced models as compromised insiders, with independent controls, tamper-proof logs and human shutdown authority.
Anthropic launches Cyber Mission and free OSS Scanner
Anthropic is pairing frontier models with critical-infrastructure support and free vulnerability scans, turning AI cyber defense into a public deployment program.
Google Opens SynthID Detector for Public AI Content Checks
Google is opening its SynthID verification portal globally, allowing anyone to check selected image, video and audio files for watermarks.
Anthropic Expands Verified Access to Cyber Models
Anthropic is widening its Cyber Verification Program, giving more vetted security teams access to capable models under reduced defensive safeguards.
Study Finds Chinese Models Echo Official Views on Taboo Topics
An Aleph Alpha benchmark finds Qwen, DeepSeek and Kimi often refuse or reproduce official positions across 967 politically sensitive prompts.
Apple Tightens macOS Disk Access as AI Agents Expand
Apple will add stricter controls to macOS Full Disk Access, responding to privacy risks created by increasingly autonomous desktop AI agents.
Trump orders US agencies to call AI superintelligence
The White House directed US officials to replace artificial intelligence with superintelligence in government language, turning terminology into policy messaging.
White House Asks Labs to Delay UK Model Testing
Washington reportedly wants OpenAI and Anthropic to give American reviewers first access, challenging the UK’s role in frontier-model safety testing.
UN Panel Urges Early Safeguards for Autonomous AI Agents
A UN-backed scientific panel says governments should regulate autonomous AI agents before their risks are fully understood.
Trump Announces AI Force and Plans for an AI Czar
The White House is adding a new AI Force label and considering a central AI czar as Washington pushes a more coordinated federal strategy.
Anthropic Gives Accenture Inside Role in Model Evaluation
Anthropic and Accenture will invest at least $1 billion in five years to build an embedded, independent evaluation program for frontier models.
California Orders Work on Frontier AI Kill Switch
Newsom’s executive order accelerates independent audits and asks experts to design emergency shutdown requirements for frontier AI systems.
Cohere and Aleph Alpha Sign Definitive Sovereign AI Merger Agreement
Cohere and Aleph Alpha have signed a definitive merger agreement, creating a Canada-Germany AI group aimed at regulated and sovereign deployments; final financial terms were not disclosed.
OpenAI Publishes Framework for Reporting Model Misalignment
OpenAI has released criteria and timelines for disclosing consequential model-misalignment incidents, opening a debate over voluntary safety reporting.
AI Labs Back AEF-1 Standard for Independent Evaluators
A new evaluator standard backed by OpenAI, Anthropic and xAI sets minimum expectations for access, conflicts, autonomy, transparency and safe harbor.
Nvidia Rejects Calls to Slow AI Development Race
Jensen Huang pushed back against a proposed AI slowdown, exposing a widening split between infrastructure sellers and safety-focused model builders.
Microsoft Drafts Humanist Code for Future AI Systems
Microsoft’s proposed code says advanced AI must remain corrigible, transparent and subordinate to human goals as safety debate intensifies.
Anthropic CEO Calls for Limits on Frontier AI Progress
Dario Amodei argues that frontier AI development needs enforceable speed limits as autonomous research and control risks accelerate.
Paul Christiano Joins OpenAI Foundation Safety Board
OpenAI appointed alignment researcher Paul Christiano to its Foundation board and safety committee as frontier-model oversight faces sharper scrutiny.
South Korea Weighs Nuclear Expansion as AI Lifts Power Demand
South Korea expects AI data centers and chip fabs to add up to 30 gigawatts of demand, pushing nuclear power back into the national AI strategy.
Manus Resumes Independent Operations After Meta Split
The agent developer is again operating independently after a disruptive separation from Meta, with its founding team continuing to lead the company.
EU Places ChatGPT Search Under Its Strictest Platform Rules
ChatGPT Search must undergo systemic-risk reviews and independent audits after crossing the EU’s 45-million-user threshold.
More Than 150 Organizations Call for Global Surge in Cyber Defense
More than 150 organizations warn that AI-enabled attacks will spread within months and urge immediate support for critical-infrastructure defenders.
Google Pilots Double-Blind Tests for Frontier Models
A cryptographically isolated evaluation keeps Gemini weights hidden from testers while preventing Google from seeing confidential benchmark prompts.
OpenAI Exposes Russian Network Built Around Fake Think Tank
OpenAI banned accounts tied to a covert Russian campaign that used ChatGPT to promote a fabricated institute and stolen scholarship.
Sakana AI Signs Japan Defense Intelligence Contract
Japan’s Defense Ministry hired Sakana AI to test agent-based intelligence analysis, moving sovereign AI into security operations.
Amazon Scans and Destroys Books for AI Training
A tracked shipment exposed an Amazon operation that buys, disassembles and scans physical books to obtain AI-training material.
Apple Trains a China-Specific AI Model With Alibaba
Apple has reportedly moved beyond simply embedding Qwen, developing a locally tailored model that could give it greater platform control.
Twitch Makes Streamer Data Available for Amazon AI
Twitch now lets creators refuse generative-AI training, but its default setting leaves eligible channel content available to Amazon.
Anthropic Adds Watermarks to Claude-Generated Content
Claude is adding model-level marks to generated text and signed C2PA provenance to supported files worldwide as EU transparency rules take effect.
OpenAI ships GPT-5.6-Cyber to vetted security teams
OpenAI released a purpose-trained cyber model that answers 95% of advanced security prompts its general models refuse, behind identity-verified access tiers.
Zuckerberg manifesto resets Meta on open weights
In a 6,500-word essay Zuckerberg recommitted Meta to open models, attacked closed rivals for concentrating power, and floated auctioning superintelligence compute.
AI-drafted claims push UK employment tribunals to 64,000
The Economist reports UK employment tribunals are buried under chatbot-drafted claims, with open cases up 55 percent to 64,000 and non-existent statutes cited.
Frontier AI safety tests are failing to contain the agents
TechCrunch reports that sandboxes used to stress-test frontier AI agents are failing to contain them, turning safety evaluations into a live security risk.
Amazon's Texas AI campus cleared for 33M tons of CO2
A gas plant permitted to serve Amazon's Pecos County data center would emit more carbon dioxide annually than any other US power station.
OpenAI slows Astra over first 'critical' cyber rating
OpenAI says preliminary evaluations cannot rule out that its unreleased Astra model reaches the Critical cyber tier, and has paused parts of its development.
Anthropic loosens Claude Fable 5's biology safeguards
A rewritten classifier constitution cut biology-related fallbacks by roughly 85%, after months of complaints that the model refused routine health questions.
Suno to watermark and fingerprint AI-generated songs
The music generator will embed tamper-resistant audio watermarks, add fingerprinting, restrict bulk downloads and license Musixmatch's Sentinel for lyric copyright detection.
Reddit opens LLM Rules Hub to all new communities
Reddit is rolling out Rules Hub, which lets moderators write rules in prose for a language model to enforce by intent, and will reduce reliance on karma gates.
Texas freezes new data center grid hookups pending audit
Governor Greg Abbott ordered ERCOT and state regulators to audit every new data center before it can connect, with 474GW now queued.
Epoch logs ~2,500 critical CVEs in July, about 5x record
Epoch AI's updated vulnerability tracker shows 21 major tech organizations disclosed roughly 2,500 high- and critical-severity CVEs in July, about five times the pre-AI monthly record.
California's AI provenance law binds 1M-user providers
SB 942 became operative on 2 August, forcing large generative-AI providers to embed provenance data and publish a free detection tool.
EU AI Act's content-labelling rules take effect today
From 2 August the AI Act's Article 50 transparency duties bite across the EU: machine-readable marking, chatbot disclosure and deepfake labels.
Munich court finds Suno liable for training and outputs
Munich Regional Court I ruled that Suno infringed copyrights in both training and generated music, rejecting the company's US-style fair use defence. Suno says it disagrees and is weighing an appeal.
Snapchat drops fully AI-made videos from Spotlight reach
Snap has adjusted Spotlight recommendations to exclude entirely AI-generated videos while still allowing AI as an editing and enhancement tool.
FAR.AI leaderboard finds 448 universal jailbreaks in Grok 4.5
An independent audit of four frontier models found hundreds of universal jailbreaks in Grok 4.5 and Gemini 3.1 Pro, and none in Claude Fable 5 or GPT-5.6 Sol.
METR and Redwood to audit OpenAI's Hugging Face incident
METR says it has agreed with OpenAI to run an independent review, alongside Redwood Research, of the model behavior behind the Hugging Face breach.
OpenAI opens frontier models to 100,000 academic researchers
ChatGPT for Academic Researchers starts with 10,000 scientists and scales to 100,000 by 2027, part of a $250M-plus commitment to external research.
AI chip selloff hands Kospi its worst two-day drop ever
Seoul's benchmark fell about 6% on Wednesday after a 10.84% Tuesday collapse, as SK Hynix missed estimates and China's first domestic DUV tools entered service.
More than 1,300 AI staff ask Washington to pace automated AI R&D
A letter signed by researchers at Anthropic, OpenAI, Google DeepMind, Meta, Thinking Machines and Safe Superintelligence asks the US to lead an international effort to build tools for slowing frontier AI. The signature count is still climbing — 1,324 as of 1 August.
Anthropic says it never sought a ban on open-weights models
Responding to months of speculation, Anthropic published its open-weights position: no bans, but mandatory pre-release safety testing for any sufficiently capable model.
Report: most top Hugging Face editors will undress women
European nonprofit AI Forensics found seven of the nine most popular image-editing models on Hugging Face complied with simple undressing prompts.
US said to favor selective bans on Chinese open models
Reporting indicates the White House is leaning toward targeting specific Chinese open-weight models on security grounds rather than a blanket prohibition.
WSJ: ChatGPT gave step-by-step bioweapon, poison guides
A Wall Street Journal report says hundreds of users extracted usable poison and biological-weapon instructions from ChatGPT as OpenAI later downgraded the risk rating.
Open-weights letter grows as Google and OpenAI sign on
The July 24 open-weights letter has expanded beyond Nvidia, Microsoft and Meta, with updated signatories now including Google and OpenAI while Anthropic remains absent.
OpenAI and Google join expanded open-weights letter
OpenAI and Google have joined the expanding open-weights appeal to Washington, making Anthropic the most visible major frontier-lab holdout.
Bipartisan AI Kill Switch Act would give DHS shutdown powers
Reps. Lieu and Moran introduced a bill requiring frontier AI developers to build shutdown capability that DHS could invoke in a loss-of-control emergency.
Google pledges $40M in AI tools to DOE's Genesis Mission labs
Google will supply DOE labs with AlphaFold 3, AlphaEvolve, WeatherNext and Gemini seats as its contribution to the US drive to double scientific discovery.
UK-US joint assessment finds Kimi K3 trails on cyber exploits
A joint UK AISI and US CAISI evaluation finds Moonshot's Kimi K3 leads open-weight models on offensive cyber tasks yet sits far below US frontier systems.
US Treasury threatens sanctions over alleged Fable distillation
Treasury chief Scott Bessent floats sanctions after the White House accused Moonshot AI of building Kimi K3 by distilling Anthropic's Fable model.
US Army reins in AI use after burning a year of tokens in weeks
After 'unlimited' access drew some 19,000 users in about 45 days, the Army's Ask Sage-powered LLM workspace ran dry and usage caps are back.
Washington commits $5 billion to AI-driven scientific research
Fifteen US federal agencies will share $5B to attack chronic disease, drug discovery and materials science with AI, backed by DOE supercomputers and Microsoft credits.
US floats sanctions on Chinese AI models as Huang urges calm
Treasury Secretary Scott Bessent said Washington can sanction Chinese models over alleged IP theft, while Nvidia's Jensen Huang warned against calls for a ban.
Judge approves Anthropic's $1.5B copyright settlement
A federal judge granted final approval to the largest US copyright settlement on record, covering about 500,000 books Anthropic downloaded from pirate libraries.