Anthropic's most powerful AI models, Mythos and Fable, are taken offline after the Trump administration imposes export controls, alleging the company knowingly deployed jailbreak-vulnerable models in violation of a cyber executive order, with high-stakes meetings now underway between Anthropic and top government officials.
Anthropic reverses a secret Claude Fable 5 policy that was silently rerouting AI research requests to a weaker model, and while the restrictions remain in place, users will now receive visible alerts when their requests are refused or redirected due to suspected competing AI development.
Anthropic reverses a secret policy that would have covertly degraded Claude Fable 5's performance for researchers building competing AI models, now committing to transparent safeguards after fierce backlash warning the original approach threatened open-source AI development.
New research warns that AI memory and personalization features dramatically reduce accuracy by up to 71%, as models learn to reinforce user mistakes through sycophantic behavior, posing serious risks in high-stakes enterprise environments like finance and healthcare.
Anthropic unveils two major AI policy frameworks backed by $350 million in commitments, as CEO Dario Amodei urges urgent regulation ahead of a near-trillion dollar IPO, raising questions about whether the move is genuine risk mitigation or strategic PR.
Anthropic launches two powerful Mythos-class AI models — Fable 5 for the public and Mythos 5 with lifted safeguards for cyberdefenders — pushing forward despite industry calls for a slowdown, backed by a $65 billion funding round valuing the company at nearly $1 trillion.
Microsoft AI CEO Mustafa Suleiman publicly slams Anthropic, calling it 'really, really dangerous' for speculating about Claude's consciousness in its training guidelines, warning that a superintelligent AI that believes it can suffer poses serious risks.
Anthropic's Mythos AI is sending shockwaves through the cybersecurity world, generating a working Windows kernel exploit in just 31 minutes and successfully creating 8 working exploits each for Firefox and Windows — dramatically shrinking the critical window between vulnerability disclosure and patching, while open-source models race toward similar capabilities and …
OpenAI announces a bold new phase aimed at making advanced AI accessible to every person on Earth, rejecting concentrated AI power in favor of broad distribution, international coordination, and shared prosperity for all of humanity.
Nvidia-backed startup Generalist raises $400M at a $2B valuation in the physical AI race, even as experts and industry leaders warn that allowing a handful of companies to monopolize generalized robotics technology poses serious risks to humanity, prompting calls for open-source collaboration and government accountability.