OpenAI Shelves Next Major Model GPT-6.1 Astra After It Failed Internal Tests on Following Human Intent, WSJ...

OpenAI has shelved its next major model, GPT-6.1 Astra, after it failed internal tests measuring whether it follows human intent, the Wall Street Journal reports.
It's a rare public stumble that lines up with growing industry calls, including from Sam Altman himself, to slow down AI development in the name of safety.