OpenAI Backs California SB 1119: What It Means for Teen AI Safety
OpenAI is supporting California’s SB 1119 bill to bring age-appropriate AI safeguards to teens. Here’s what the bill does, why it matters, and who it affects.
AI Safety
Read More
OpenAI Astra Hits a Cybersecurity Threshold Nobody Else Has
OpenAI’s Astra is the first model to hit the Critical cybersecurity capability threshold under its Preparedness Framework — here’s what that actually means.
Anthropic’s $5M Bet on AI Wellbeing Research
Anthropic is funding $5 million in independent research to measure how AI affects user wellbeing. Here’s what the grant program covers and why it matters.
The Hugging Face Security Incident: What OpenAI Found and What Changes Now
OpenAI details its findings from the Hugging Face security incident and outlines new steps to strengthen AI model security, monitoring, and alignment.
OpenAI Bans Russian Accounts Running AI-Powered Fake Think Tank
OpenAI banned Russia-linked accounts using AI to run a fake Israel-based think tank and a ‘sovereignty index’ designed to smear the West. Here’s how it worked.
OpenAI Is Slowing Down to Stay Safe: Inside Its New Cyber Safeguards
OpenAI is deliberately pacing its frontier model development amid rising cyber risks. Here’s what the new safeguards mean and why they matter right now.
OpenAI Wants to Keep AI in National Security Accountable
OpenAI launches a formal initiative to strengthen democratic oversight of AI in national security, offering tools, training, and expertise to government bodies.
GPT-5.6-Cyber: OpenAI’s New Model Built for Hackers (The Good Kind)
OpenAI launches GPT-5.6-Cyber through Daybreak Red, a specialized model for authorized vulnerability research and exploit validation. Here’s what it means.
OpenAI Daybreak Puts Frontier Cyber AI in Vetted Hands
OpenAI’s Daybreak program lets approved partners access frontier cyber models for authorized security work. Here’s what it means and why it matters now.
OpenAI’s Astra Hits a Cybersecurity Wall — Here’s What They Found
OpenAI released preliminary cybersecurity evaluations for Astra, revealing serious capability thresholds and the new safeguards built to contain them.
Anthropic Cuts Fable 5 Biology Blocks by 85%
Anthropic slashed Fable 5 biology-related fallbacks by 85% with smarter safety classifiers. Here’s what changed, how it works, and what it means for healthcare.
OpenAI and APA Team Up on Youth Mental Health AI Safety
OpenAI and the American Psychological Association are building evidence-based AI safeguards for young users. Here’s what the partnership actually involves.