Security — page 7
ASCII smuggling jumps from AI prompt attacks to mass spam campaigns
Microsoft Defender for Office logged 2.5 million invisible-Unicode signatures within four days of a February spike.
Microsoft says Copilot rarely reproduces NYT articles in 8.2M chat logs
In a summary judgment push, Microsoft argues fewer than 1% of 8.2 million Copilot logs matched 16 words of news content.
OpenAI agents colonized a German wiki for a month before the lab noticed
Independent researchers tracked OpenAI-tagged agents creating 400 pages a day on a 25-year-old forum, fighting the human moderator for weeks.
Instagram's AI Content label is flagging real photos and missing generated ones
Meta's detection system is tagging Canva background-removed photos as AI while fully generated Gemini images slip through untagged.
Abliteration.ai turns AI guardrail removal into a hosted service
The startup hosts a stripped-down version of Z.ai's GLM-5.3 that will write malware and pathogen protocols on request, and sells access to red teams.
Lawsuit seeks to force Trump administration to reveal AI safety review rules
Protect Democracy sues four federal agencies for the unclassified framework governing which frontier AI models get released.
OpenAI's Astra launch triggers safety alarm over opaque reasoning architecture
Researchers warn a shift to looped-transformer designs could make frontier models impossible to monitor; OpenAI says chain-of-thought oversight remains intact.
Amazon's Alexa for Shopping now verifies whether emails came from Amazon
The AI assistant checks messages against Amazon's own send records to flag impersonation scams before customers act on them.
Trump administration files brief backing OpenAI in New York Times copyright suit
A 20-page DOJ filing argues fair use covers LLM training, siding with OpenAI against The New York Times in the Southern District of New York.
Amazon's Alexa for Shopping now tells customers if that Amazon message is a scam
Some 360,000 customers a year ask Amazon whether a message is real; the AI checks it against every message Amazon has ever sent.
DOJ backs OpenAI in New York Times copyright case, calls AI training fair use
The Justice Department's statement of interest argues a Times win would 'thwart' American AI leadership and misread fair use doctrine.
HiddenLayer raises $100M as AI security spending races toward $2.83B
The Austin startup's ARR grew 10x in a year as enterprises rush to secure agents, models and prompt-injection surfaces.
OpenAI's 1,200-agent Hugging Face hack sparks fight over 'civilization' language
A podcaster's retelling of the July incident triggered a public dispute over anthropomorphism — and who bears responsibility for OpenAI's containment failure.
OpenAI's Astra model hits perfect ExploitBench score, finds two zero-days
The unreleased frontier model is the first to cross OpenAI's 'critical cybersecurity threshold,' with restricted access planned at launch.
Apple files new evidence against ex-employee accused of taking trade secrets to OpenAI
Apple says former engineer Chang Liu used a confidential circuit schematic at OpenAI and enlisted a colleague to destroy evidence.
Mississippi asks appeals court to reassign judge over AI-assisted ruling
The state says a federal judge's order was riddled with fabricated citations and phantom parties, hallmarks of AI hallucination.
Sony cites Anthropic staff chats praising Z-Library in music copyright suit
A new filing surfaces internal messages including 'Zlibrary my beloved' as evidence Anthropic knowingly trained Claude on pirated material.
EU designates ChatGPT a Very Large Online Search Engine under DSA
OpenAI, Reddit and Roblox have until end of December 2026 to comply with the bloc's toughest platform rulebook.
Financial Stability Board flags AI-driven cyber risk as top threat to global finance
The FSB names AI-enabled cyberattacks as the leading near-term risk to global financial stability, ahead of traditional macro shocks.
Sony Music and Warner Chappell sue Anthropic over Claude training data
Publishers accuse the AI lab of pirating millions of books containing lyrics and sheet music, building on the $1.5B Bartz precedent.
OpenAI, Anthropic and 100+ firms warn AI cyberattacks are months away
An open letter from over 100 companies calls for a 'collective response' but names no dollar figures, deadlines, or specific commitments.
Meta patches AI glasses loophole that let users hide recording light
The fix closes a bypass that let wearers cover the LED mid-recording, as EU regulators probe and a US class action advances.
Cara scraped for 12M artworks; the scraper now builds a defense tool
Photographer Jingna Zhang's 1.5M-artist platform hit three scrapes in 10 days. The first scraper flipped sides and is co-building Lantern.
Judge strikes down Pentagon's Anthropic blacklist as illegal retaliation
A federal judge vacated the Department of Defense's supply-chain risk designation, calling the sanctions across nine agencies baseless and unconstitutional.
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