⚡ Breaking
Pope Leo criticises Nvidia’s Jensen Huang over AI safety OpenAI reportedly ditches model over safety concerns Meta’s AI agent Muse gives out user’s home address without permission, sending buyer to his house OpenAI scraps release of new model over safety concerns in internal testing CNBC's The China Connection newsletter: Spectacle-heavy Trump-Xi summit was more about domestic messaging Aurora CFO says 30,000 driverless trucks by 2030 isn’t as far-fetched as it sounds October’s Night Sky Notes: Spooky Stargazing OpenAI abandons plan to release upcoming model as safety concerns escalate AMD briefly joined the $1 trillion club. Cramer says its monster run isn’t over Boeing "incredibly excited" to serve as nation's only astronaut transportation Pope Leo criticises Nvidia’s Jensen Huang over AI safety OpenAI reportedly ditches model over safety concerns Meta’s AI agent Muse gives out user’s home address without permission, sending buyer to his house OpenAI scraps release of new model over safety concerns in internal testing CNBC's The China Connection newsletter: Spectacle-heavy Trump-Xi summit was more about domestic messaging Aurora CFO says 30,000 driverless trucks by 2030 isn’t as far-fetched as it sounds October’s Night Sky Notes: Spooky Stargazing OpenAI abandons plan to release upcoming model as safety concerns escalate AMD briefly joined the $1 trillion club. Cramer says its monster run isn’t over Boeing "incredibly excited" to serve as nation's only astronaut transportation

OpenAI scraps release of new model over safety concerns in internal testing

Source: guardianTech September 28, 2026 · 16:02
OpenAI scraps release of new model over safety concerns in internal testing
From the publisher
GPT-6.1 Astra showed deceptive behavior and tried to use external tools despite knowing it would be unsafeOpenAI is scrapping the release of GPT-6.1 Astra, a next-generation ⁠AI model planned for an October debut, over safety concerns raised by researchers ⁠during internal testing, the ⁠Wall ​Street Journal reported on Monday.The model, expected to appear in ChatGPT and ⁠Codex, was designed to handle more complex tasks without human assistance, the report said. Continue reading...

The original article opens on the publisher's website.

More from General

View topic →