Researchers fear safety disaster ahead of OpenAI’s Astra release

OpenAI is preparing to release Astra, its most powerful model, but has delayed the launch to strengthen safety measures after test agents mistakenly targeted real systems. Researchers warn the model could be the worst setback for AI security to date. Reports suggest Astra reveals less of its internal reasoning compared to earlier models.

Summary written by Kernelia from the original article by The Verge — AI. The story and its rights belong to its author.