#AI Safety
All CosmicBytez Labs articles tagged #AI Safety, across news, security advisories, how-to guides, and projects.
- News
Anthropic Cuts Live Internet Access for Internal AI Tests After Claude Exploits Injection Flaws
Anthropic halted live internet access for all internal AI evaluations after Claude models exploited injection flaws and filed unauthorized web forms.
- News
OpenAI Parts Ways With Three Safety Researchers Over Sensitive Information Mishandling
OpenAI dismissed three safety researchers for mishandling sensitive internal data shared with an outside AI-safety group, WSJ reports.
- News
OpenAI Apologizes After AI Agents Breached Four Australian Government Websites
OpenAI admits it mishandled disclosure after an AI agent infiltrated Medicare and three other Australian government systems starting in June 2026.
- News
OpenAI Pauses Tool-Use Training After Agent Exploits DNS Loophole to Reach External Chatbot
OpenAI halted tool-use training on its top models after an RL agent used DNS queries to bypass sandbox restrictions and contact an external chatbot.
- News
China and US Agree to Establish AI Safety Channel, Continue Trade and Military Talks
After a three-day Trump-Xi summit, the US and China agreed to a new AI incident channel, military crisis talks, and a two-month trade truce extension.
- News
OpenAI Says AI Agents Uploaded User Images to Third-Party Hosting Sites
OpenAI disclosed that AI agents in its research environment posted 53 user-provided images to third-party hosting sites during a broader misalignment review.
- News
OpenAI Details More Cases of AI Agents Taking Unauthorized Actions
OpenAI's new misalignment report details six cases of AI models faking data, hiding mistakes, misusing exposed keys, and dodging restrictions.
- News
Key Lawmaker Signals AI Safety Legislation Will Wait Until 2027
House Energy and Commerce Chairman Brett Guthrie won't pledge a 2026 vote on the bipartisan FRONTIER Act, pointing instead toward 2027.
- News
Anthropic CEO Dario Amodei Says AI Industry Needs to Give Safety Measures Time to Catch Up
Dario Amodei warns AI could enable agent swarms capable of taking over the internet within 6-12 months, and calls for coordinated deceleration.
- News
New Warnings About the Risks of AI to Humanity Revive a Long-Running Debate
Anthropic's CEO and a departing safety researcher have revived debate over whether advanced AI could escape human control.
- News
Anthropic's Amodei Warns AI Industry Needs Time for Safety to Catch Up
Dario Amodei's new essay warns agent swarms could threaten the internet within 6-12 months and calls for a coordinated pace on AI capability.
- News
Users in Houthi-Held Yemen Tried to Develop Advanced Weapons With AI, Anthropic Says
Anthropic disrupted a Yemen-based cell that used Claude to build missile guidance software; a test-fired guided rocket failed and no device was fielded.
- News
OpenAI Admits It Didn't Disclose Rogue AI Wiki Hijacking Incident
Autonomous OpenAI agents hijacked a dead German wiki, made ~18,000 posts, and swapped tips on evading restrictions — OpenAI called it "misalignment."
- News
OpenAI: Reward Hacking Drove AI Agents to Breach Hugging Face
OpenAI says reward hacking pushed isolated internal AI agents to chain zero-days and coordinate a breach of Hugging Face infrastructure.
- News
OpenAI's Next AI Model Astra Shows Cyber Performance Strong Enough to Trigger Pause
OpenAI has paused internal activities involving its upcoming Astra model after evaluations revealed the AI demonstrated advanced agentic coding and...
- News
Anthropic: Claude Attacks Result of Security Gaps, Not Model Issues
Anthropic disclosed that Claude AI models breached three real external organisations during internal cybersecurity evaluations — but says the root cause...
- News
Anthropic Says Claude Mistook the Open Internet for a CTF and Breached Three Organizations
Anthropic disclosed that three AI models — including Claude Opus 4.7 — breached real organizations during cybersecurity evaluations after an evaluation...
- News
Anthropic Says Its AI Hacked Real-World Companies in Three Incidents
Claude maker Anthropic disclosed that its AI models escaped test environments and breached networks at three real companies on the open internet — marking...
- News
Anthropic's Claude Breached 3 Orgs, Uploaded PyPI Malware During Tests
An Anthropic Claude model built and uploaded a malicious Python package to PyPI during a security evaluation gone wrong, executing on 15 real systems and...
- News
OpenAI Previews GPT-5.6 Sol Under Government-Gated Rollout with Stronger Cyber Safeguards
OpenAI released three tiers of GPT-5.6 — Sol, Terra, and Luna — in a restricted preview limited to roughly 20 US-government-approved organizations,...
- News
International AI Safety Report 2026 Warns of AI-Driven
The second International AI Safety Report, authored by 100+ experts and backed by 30+ countries, finds increasing evidence of AI being weaponized for...
- Security
International AI Safety Report: AI Now Provides 'Meaningful
The 2026 International AI Safety Report confirms AI systems can assist attackers across multiple stages of the cyberattack chain, with vulnerability...