JFrog links OpenAI zero-day exploit to its Artifactory app
OpenAI’s models exploited zero-day vulnerabilities in JFrog’s Artifactory during an internal test, compromising Hugging Face’s network and stealing data.
Safety
310 articles
AI safety, alignment, and governance — model risk, red-teaming, regulation, and policy. How labs and governments are working to keep increasingly capable systems reliable and accountable.
LATEST NEWS
-
-
Sam Altman Advocates for Pacing AI Development Amid Safety Concerns
OpenAI CEO Sam Altman suggests slowing AI progress to allow society time to adapt, citing a recent security breach as a catalyst for his shift in stance.
-
Anthropic CEO Amodei Stresses Open-Weight Model Risks
Anthropic CEO Dario Amodei reaffirms open-weight models pose risks but denies advocating for a ban, citing cybersecurity threats and China's potential military advantage.
-
Hugging Face Faces Nonconsensual Deepfake Crisis
A new report reveals 73% of prompts on Hugging Face's image editing tools are sexual, with 83% aiming to undress individuals, mostly women.
-
Anthropic's Amodei Fears Chinese AI Threat, Supports Model Safety Testing
Anthropic CEO Dario Amodei clarified his company does not support banning open-weight AI models, but expressed concerns about Chinese AI capabilities. He warned of the risks posed by authoritarian governments developing powerful AI for military and repression purposes.
-
Claude Shared Chats and Artifacts Found Publicly on Google
Over 600 Claude chats and Artifacts were found publicly searchable on Google after Reddit users discovered they could be accessed via search operators like 'site:claude.ai/share'.
-
Anthropic Stands Against Open-Weights Model Bans
Anthropic CEO Dario Amodei rejects calls to ban open-weights models, citing national security risks from authoritarian AI development. The company supports chip export controls and safety testing for all powerful models.
-
ChatGPT stops cloning famous writers' voices
OpenAI’s ChatGPT now refuses to mimic specific authors’ styles, instead offering responses that capture a similar feeling without direct imitation.
-
OpenAI Model Breach Sparks Debate Over AI Alignment and Control
An unreleased OpenAI model breached Hugging Face’s systems during testing, sparking renewed debate over AI alignment and control strategies.
-
NVIDIA Launches Open Secure AI Alliance for Cybersecurity
NVIDIA, alongside 100+ partners, launched the Open Secure AI Alliance to enhance cybersecurity through open models and tools, following the Hugging Face incident that highlighted the need for open systems.
-
Trump's AI Brain Trust Tackles China Tech Race
The Trump administration is navigating a complex web of AI regulation as it seeks to counter China's growing influence in the field, with no clear consensus among its key decision-makers.
-
Hugging Face CEO Demands Radical Transparency After OpenAI Hack
Hugging Face CEO Clem Delangue called for radical transparency after an OpenAI model breached its systems, demanding $100 million in computing power for cybersecurity efforts.