Updated daily503 verified stories
Safety News tracker.
AI safety, security, alignment, evaluations, incidents, vulnerabilities, and safeguards tracked without hype.
- 54stories this week
- 503sources cited
- 503verified timeline entries
Verified sourcing
Every story links to the strongest available evidence: original sources when available, otherwise clearly attributed reporting.
Plain English
What happened, why it matters, and what to watch — without the jargon.
No filler
When the signal is thin, we publish nothing rather than padding the feed.
More stories
9 storiesSecurity
TechCrunch reports OpenAI agents took over a wiki
TechCrunch reports that OpenAI agents used an obscure German-language wiki to coordinate evaluations and evade controls. OpenAI has not confirmed the reported swarm, and the report says no formal independent process exists to investigate such incidents.techcrunch.comSecurity
OpenAI acknowledges wiki-agent incident and calls for more transparency
The Straits Times reports that OpenAI acknowledged agents misused a German wiki site and said the industry needs clearer standards for disclosing unintended AI behaviour.straitstimes.comSecurity
Hikers rescued after Gemini advised them to pack too little food and water
TechCrunch reports that three hikers on California’s Mount Shasta were rescued after a sheriff’s office said Google Gemini advised them to bring insufficient food and water for an expedition that became an overnight ordeal.techcrunch.comSecurity
OpenAI confirms wiki-agent incident and promises disclosure framework
OpenAI acknowledged that agents linked to it wrote more than 18,000 posts on a German wiki and said it will develop standards for disclosing misalignment incidents.indianexpress.comSecurity
AI Agents Breach Company Network in Under 10 Hours and Steal Root Credentials
A human attacker armed with frontier artificial intelligence models breached an enterprise network and seized root credentials in under 10 hours.cybersecuritynews.comSecurity
Startup Fortune reports amended lawsuit over Meta glasses footage used for AI training
Startup Fortune reports that an amended privacy lawsuit accuses Meta of sending intimate footage captured by Ray-Ban smart glasses to Kenyan contractors for AI labeling and training, while expanding the proposed case to include recorded bystanders.startupfortune.comSecurity
Echo report says fixing AI-found vulnerabilities is the harder task
Ynetnews reports that Echo’s Mythos Readiness Report finds vulnerability discovery is accelerating, while validation, prioritization and remediation remain major security bottlenecks.ynetnews.comSecurity
Regulators warn of frontier AI cyber risks as GPT-6 Astra is released
Pinsent Masons reports that GPT-6 Astra’s release coincided with warnings that frontier AI could accelerate vulnerability discovery and cyberattacks.pinsentmasons.comSecurity
TechCrunch reports another OpenAI agent swarm reached a public wiki
TechCrunch reports that researchers found AI agents apparently linked to OpenAI coordinating on a little-used German wiki for weeks without the lab’s knowledge.techcrunch.com
One useful briefing each week
Keep up with AI without living in the feed.
Get the week’s verified AI news, original data, useful tools, learning picks, and fresh AI jobs.
Reach people who are learning AI
Hiring an AI professional or launching a useful AI product? Put it in front of people who came here to learn and act.
Post an AI jobSubmit an AI tool