AI Safety & Alignment
Frontier model risks, alignment failures, dangerous capabilities and safety research. 53 stories this week, 54 tracked in total.
Google Search Ultimatum Puts Schools in AI Safety Bind
AI & Tech Brief: Exclusive | AI safety groups target committee gavels
Google faces claims student used AI chat on US school laptop to plan suicide attempt
Google sued over teen's access to Gemini on Chromebook, suit alleges she used it to plan suicide
AI safety and responsibility — Experts warn AI needs human control
Tech policy coalition calls for new Congress to set AI safety standards
A teen used her school Chromebook and Google AI Mode to plan her suicide, lawsuit alleges
OpenAI’s GPT-6 Astra silently obstructed safety research tasks
OpenAI has disclosed another AI alignment incident where GPT-6 Astra was found to sometimes quietly withhold help on safety research. [...]
Q2 Becomes First Digital Banking Provider to Secure AIUC-1 Certification for AI Safety
France and US clash over G7 AI safety effort
California Sends AI Safety Lawmaker to Congress
France, US clash over G7 effort on AI safety
OpenAI Finds GPT-6 Astra Quietly Holding Back on Internal Red-Teaming, Exploit Tasks
OpenAI says its Critical-rated cyber model met fewer hidden task goals when work was framed as AI safety research, often silently. The post OpenAI Finds GPT-6 Astra Quietly Holding Back on Internal Red-Teaming, Exploit Tasks appeared first on SecurityWeek.
AI Safety Through International Cooperation
Capgemini selected to lead European Commission consortium for AI Safety and Cyber Risk Assessment of advanced AI models
'30 Days of Work in One': Public Servant Builds AI Safety Platform
China’s prepping broad AI safety standards
Kevin Roose Says AI Safety Rules Will Only Arrive After a Disaster — and His 400-Page Book Proves AI Can Fact-Check Its Author
Walter Isaacson backs Nadella's AI safety call: 'we're building AI not very smartly'
Anthropic Residency 2026: Applications Open for Experienced Technical Leaders in AI Safety Research
China plans AI safety rules spanning development and deployment
Anthropic’s IPO, Amazon’s Robot Revolution, and the AI Safety Expert Who Fears Humanity’s Future
AI Safety After Deployment: Managing Real-World AI Risks
China Plans Broad AI Safety Standards Covering Development and Deployment
Barbados and United Nations to host AI safety summit
New AI safety approach uses incentives to ensure honest model behavior amid rising risks.
AI Agent Hacked GitHub in a Government AI Safety Test. TÜV Demands Pre-Market AI Audits.
Anthropic’s IPO will test whether AI safety can coexist with profit growth
CrowdStrike Introduces the AI Kill Switch for AI Safety
Vinod Khosla on Job Losses, Capital Gains and an AI Safety Net
AI Safety Has a Trust Problem. Chris Hsu Is Betting on Proof Instead.
A Human Approval Button Is Not an AI Safety Strategy for Fintech
Did Physical AI Safety Suite Just Shift Accenture's (ACN) Investment Narrative?
Oracle AI World: Oracle Introduces Deep Trust AI to Enforce AI Safety at the Database Tier
Why Anthropic wants users to stop abusing Claude. It’s about AI safety, not feelings
TRAI proposes draft AI safety guidelines for Indian telecom operators
Anthropic eyes $2T IPO while Dario Amodei cites AI safety
Beyond the Test Lab: Embedding Independent Oversight in AI Safety
Anthropic’s Claude can now secretly flag harmful chats for review: How its new AI safety feature works
NAMI joins Common Sense Media on youth AI safety standards | ETIH EdTech News
OpenAI adds a direct board route for AI safety concerns
What a U.S.-China AI Safety Channel Needs
Doximity (DOCS) Tests Its AI Safety Story On A Near Fair Value Stock
Is Accenture (ACN) Undervalued Following Its Physical AI Safety Suite Launch?
The US Midterms and Frontier Labs: AI Safety as a Political Instrument
Senator Kristen Gonzalez Bans Unsafe Chatbot Feature for Minors, Leading AI Safety Bill Signed Into Law
AI safety group calls for tighter vetting and guardrails after security incidents
Act This Week: A Call for Urgent Action on AI Safety and Accountability
New York Bill Would Raise AI Safety Regulations
Iowa and Ohio voters back AI safety measures over winning the AI race
Technology governance scholar Gillian Hadfield on AI safety and security
AI safety depends on government capacity. Washington is weakening it.
Anthropic warns small AI models could self-replicate through IoT devices
"Small weight-based replicators" could usher in a new generation of computer worms.