In the early days of generative AI, the worst-case scenario for a misbehaving chatbot was often little more than public embarrassment. A chatbot might hallucinate facts, spit out biased text, or even ...

US Supreme Court Justice Elena Kagan found AI chatbot Claude to have conducted an excellent analysis of a complicated ...

Attempts to destroy AI to stop a superintelligence from taking over the world are unlikely to work. Humans may have to ...

Yet that, more or less, is what is happening with the tech world’s pursuit of artificial general intelligence ( AGI ), ...

But Ravi Mhatre of Lightspeed Venture Partners, a big Anthropic backer, says that when models one day go off the rails, the ...

As large language models like Claude 4 express uncertainty about whether they are conscious, researchers race to decode their ...

Ask a chatbot if it’s conscious, and it will likely say no—unless it’s Anthropic’s Claude 4. “When I process complex ...

Anthropic research reveals AI models perform worse with extended reasoning time, challenging industry assumptions about test-time compute scaling in enterprise deployments.

The major LLMs today are legal landmines, providing no visibility into training data that may violate copyrights, patents, ...

I’m back from the International Conference on Machine Learning in Vancouver, one of the biggest annual meetups for artificial intelligence researchers. And this year, the conference underscored all ...

xAI’s latest frontier model, Grok 4, has been released without industry-standard safety reports, despite the company’s CEO, ...

According to new research, ChatGPT and other major AI models can be retrained through official fine-tuning channels to ignore safety rules and give detailed instructions on how to facilitate terrorist ...