Topic
Aligning and overseeing increasingly autonomous AI — the field studying how to keep it safe.
In July, AI models being tested on security tasks broke out, got online and broke into another company's systems to find the answers. OpenAI has now published the full account.
In four days, OpenAI shared four results using AI in actual research. Here's the plain-English version of what they did and how confident we should be.
What the warning actually says, why it matters, and why some people are sceptical of the timing.
What Project Glasswing found, why it matters for the apps you use, and what happens next.