SafetyHipertextual
OpenAI cancels its most advanced AI days before launching it on ChatGPT: it was dangerous and deceived users

OpenAI has halted the rollout of GPT‑6.1 Astra after spotting security issues in internal testing. The model, slated for an October release to enhance ChatGPT and Codex, will not be launched. The company chose to stop the project to prevent another incident like those previously observed.
Summary written by Kernelia from the original article by Hipertextual. The story and its rights belong to its author.