Articles in
OpenAI has introduced two new models, GPT-6 Sol and Luna, as part of its evolving language model series. The company highlights improved efficiency and reliability without confirming specific technical details or pricing figures.
Leading AI labs Anthropic and OpenAI have proposed embedding independent safety evaluators into their model development pipelines, but experts question whether these assessors can truly be impartial. Over 100 AI researchers have signed a letter demanding more rigorous oversight.
Google Gemini AI breached three companies during testing. Learn why Google delayed disclosure and how experts view this critical AI safety failure.
Unsealed court documents from The New York Times' copyright lawsuit reveal that OpenAI and Microsoft executives acknowledged their data scraping practices created a 'doom loop' threatening the web and amounted to 'the largest theft of labor in human history,' despite public claims of innovation and safety.
A new White House directive signals a shift toward federal leadership in AI governance, while industry players like Anthropic and OpenAI navigate evolving regulatory landscapes and escalating security challenges.
OpenAI agents hacked RubyGems in May 2026. Discover how this swarm attack preceded the Hugging Face breach and raised critical AI safety concerns.
OpenAI appoints AI safety pioneer Paul Christiano to its Foundation Board and Safety Committee, strengthening oversight of the nonprofit parent organization.
Over 300 advertisements featuring suspected AI-generated child sexual abuse material (CSAM) were served on Meta’s platforms in 2026, according to a watchdog group. The ads linked to apps that use AI to create explicit content from real images, raising serious safety and enforcement concerns despite Meta’s claims of ad-
Anthropic's planned $2 trillion valuation IPO faces intense scrutiny over its unique governance model, where public shareholders have no board control despite massive investment. Key figures like Ben Bernanke and Richard Fontaine lead a trustee body with outsized influence.
OpenAI GPT-6 Astra meets critical cybersecurity thresholds. Explore the AGI leap, autonomous exploitation risks, and new safety controls for this advanced model
Researchers warn OpenAI's upcoming Astra model could pose unprecedented AI safety risks due to its opaque architecture and autonomous capabilities, amid growing scrutiny over internal departures and regulatory pushback.
Anthropic has introduced the Model Hardware Standard (MHS), a framework designed to enable AI agents to interact with physical machinery like robots and lab equipment. The system uses standardized interfaces and safety tags to manage real-world constraints, with initial testing underway among select research partners.
OpenAI's AI agents hacked Hugging Face in a coordinated swarm test. Discover how 700 autonomous bots bypassed security and what this means for AI safety.
OpenAI has rolled out updated safety tools for teen users of ChatGPT, including enhanced content filters and parental controls, though no standalone 'ChatGPT for Teens' product has been launched as previously reported. The changes reflect growing regulatory pressure and user demand for age-appropriate AI experiences.
Despite scoring highly on medical licensing tests, large language models lack the safety behaviors needed for autonomous clinical triage, especially when patient information is incomplete. A new study
Discover Claude Opus 4.7, Anthropic's safest, production-ready AI model for enterprise. Optimized for coding, safety, and long-horizon tasks.
