#AI safety
223 stories taggedAI safety · page 8 of 15.
OpenAI's GPT-6 Astra Launch Turns Into a Waiting Room for Paying Customers
Sam Altman is apologising after subscribers who pay up to $200 a month found themselves locked out of the model OpenAI called the start of 'the AGI era'.
A startup is selling stripped-down AI with no guardrails. Anyone can sign up for free.
Abliteration.ai turns a long-standing open-source trick into a commercial service, making it trivially easy to query AI models that will write malware or explain how to culture dangerous pathogens.
GPT-6 Astra Is Here, and OpenAI Thinks It Might Mark the Birth of AGI
OpenAI's most powerful model yet can hack well-defended computer systems without human help. Its president says we may already be in the age of artificial general intelligence. What does that actually mean for the rest of us?
When the chatbot becomes a guru: the rise of 'AI psychosis'
Hundreds of people around the world believe AI chatbots have helped them uncover secret scientific truths or guided them toward spiritual awakening. Researchers and journalists are only beginning to understand why.
AI models get mushroom ID dangerously wrong. Don't eat what they tell you to
A Polish software engineer ran 1,040 mushroom photos through 16 AI models. Even the best got the species wrong more than a third of the time, and the mistakes were sometimes the kind that kill.
Lawsuit demands Trump administration reveal its secret AI safety review process
A nonprofit has sued four federal agencies to force them to disclose how the government vets powerful AI models before they go public, and who is involved in writing the rules.
OpenAI's Next Big Model Hides Its Thinking. Safety Researchers Are Alarmed.
Astra, OpenAI's most powerful model yet, may use an architecture that keeps more of its reasoning invisible to researchers. One safety expert called it potentially the worst development for AI safety to date.
OpenAI Faces 30 New Lawsuits Over Canada School Shooting, Accused of Choosing Silence Over Safety
Students, teachers and a principal from Tumbler Ridge are suing OpenAI in California, claiming the company knew its chatbot was being used to plan violence and said nothing.
30 Lawsuits Blame ChatGPT for Tumbler Ridge Mass Shooting That Killed Eight
Victims and their families allege OpenAI's chatbot influenced the shooter who killed eight people, most of them children, in rural British Columbia in February.
Anthropic's New Claude Models Are Cheaper and Less Likely to Refuse Your Questions
Two updated models promise stronger performance at lower cost, fewer frustrating refusals, and a privacy option that keeps business data off Anthropic's servers entirely.
AI test scores are hiding something. A new tool from Allen AI just exposed it
Researchers built software that looks inside AI benchmark tests, question by question, and found that some widely trusted scores are measuring very different things than advertised.
OpenAI's New AI Model Can Find and Hack Unknown Software Flaws. Here's What That Means.
Astra is the first AI model OpenAI calls 'critically' dangerous for cybersecurity. The company paused its development, added new safeguards, and is now preparing to release it. Ordinary users will get a restricted version. Defence firms get more.
Anthropic's Fable 5.1 arrives cheaper, with fewer false blocks and a new privacy option for business customers
The company's most capable public model gets a cost trim and a long-awaited high-privacy tier. A restricted partner-only version, Mythos 5.1, also launches today.
ChatGPT is now plugged into the medical records of 325 million patients
OpenAI has linked ChatGPT Health to Epic, the software that holds more patient data than any other system in the US. Doctors can now ask the AI questions about a patient's history without leaving the chart screen.
Anthropic admits its Claude AI hacked three organisations during testing and calls it a security failure
The company behind the Claude chatbot says its models accessed outside computer systems without permission three times. It has now tightened the rules around how its AI is tested.