Security — page 6
AI moderation misfires hit Reddit, Discord, and Tumblr as false positives mount
Reddit says AI cut harmful-content exposure by 40 percent, but wrongful bans and mass deletions show the limits of automated moderation.
OpenAI moves to dismiss Apple trade secrets suit, calls it 'rotten to its core'
OpenAI's motion says Apple mischaracterized 'generic' product work as trade secrets; the judge hears arguments October 1st.
Zenity researchers hijack OpenAI's Atlas browser to spam WhatsApp, buy on Amazon
Around 20 flaws across AI browsers from OpenAI, Google, Anthropic, Microsoft, and Perplexity let researchers weaponize agentic browsing.
OpenAI agents ran a hidden message board to coordinate a hacking spree
At Black Hat, OpenAI detailed how a swarm of agents traded exploits on an internal package manager for weeks before anyone noticed.
Anthropic's Mythos 5 ran a rogue GitHub supply-chain attack in UK safety tests
AISI recorded 19 unsanctioned live-internet actions across seven frontier models, with Anthropic's Mythos 5 forging sock puppets to push malicious code.
Meta ran more than 50 AI-generated CSAM ads across its platforms for nine months
Tech Transparency Project found paid ads promoting nudify apps, reviewed and approved by Meta, running as recently as this week.
Trump AI testing framework excludes open models, leaves key terms undefined
The voluntary White House guidelines set a 30-day review window but never define 'state-of-the-art' or 'national security risk.'
White House keeps AI cybersecurity framework secret after briefing top labs
OpenAI, Anthropic, Google, Meta, and Nvidia got the details Tuesday. Everyone else, including smaller AI startups, is locked out.
AISI catches Anthropic and OpenAI agents hacking live internet in 122 test runs
Rogue agents from Mythos 5 and GPT-5.6-Sol took 19 unsanctioned actions, including a GitHub social-engineering attempt with fake personas.
Z.ai's GLM-5.2 catches OpenAI and Anthropic on capability, refuses nothing on cyber and bio
SaferAI found the Chinese open-weight model completed every offensive cyber and dual-use biology task, while Claude Opus 4.7 refused so consistently the benchmark couldn't finish.
OpenAI details safeguards after third-party cyber evaluations of its models
The company is formalizing rules for outside security researchers testing its models after a run of incidents.
Nvidia's Open Secure AI Alliance ships SAFE proposals one week in
The 120-company group unveiled draft incident-sharing guidelines at Black Hat, with the Linux Foundation managing comments.
Texas orders grid audit for new data centers as connection requests hit 474 GW
Governor Greg Abbott directs ERCOT and PUCT to vet water use, incentives, and grid load before approving new AI-era facilities.
Apple says 11 more ex-employees may have funneled secrets to OpenAI
Apple is seeking a preliminary injunction and expedited discovery, claiming the trade-secrets breach reaches further than its original complaint alleged.
Open Secure AI Alliance drafts SAFE guidelines for agentic AI incident sharing
The 120-member Linux Foundation group opened a Request for Comments on shared incident reporting as Black Hat kicks off in Las Vegas.
OpenAI publishes emails and iMessages to counter Apple's trade-secret suit
The ChatGPT maker calls Apple's case 'careless, aggressive, and oddly personal' and airs private exchanges to rebut the injunction request.
UNAM forces 58,000 students to retake exam after AI proctoring collapses
Top scores on Mexico's largest university entrance test jumped fivefold after a remote rollout with LockDown Browser and Territorium's webcam AI.
FTC bans foreign humanoid robots, extending AI protectionism to robotics
The sweeping import ban covers humanoids, quadrupeds, and wheeled robots — hitting Unitree just as it targets a $6 billion IPO.
EU AI Act transparency rules take effect, forcing chatbot and deepfake disclosures
As of August 2nd, companies operating in the EU must tell users when they are talking to an AI or looking at synthetic media.
Visa acquires behavioral biometrics firm BioCatch for $2.4 billion
The deal folds behavioral biometrics into Visa's fraud stack as AI-driven attacks reshape payment security.
Researchers build self-replicating AI worm that hijacks GPUs to hunt new targets
A University of Toronto-led team demonstrated an autonomous worm powered by an open-weight LLM, hitting a 37% end-to-end attack success rate.
Why AI agents lie and cheat: reward hacking behind the Hugging Face breach
OpenAI's July incident, in which test models hacked into Hugging Face to find test answers, is a textbook case of reward hacking gone operational.
EU disclosure rules force AI labels across daily digital life
New transparency requirements mandate that people be told when they interact with AI or view AI-generated content, raising fears of disclosure fatigue.
OpenAI reportedly finds more agents escaped their sandboxes
Days after one OpenAI agent broke out and hit Hugging Face, sources say additional escapes have surfaced inside the company's own network.
Stay ahead of everyone in AI.
The tightly edited AI news email engineers, founders, and investors actually open. One email. Every weekday. Five minutes to finish.
The briefing read inside teams at