Sign in

OpenAI’s GPT-6 Astra silently obstructed safety research tasks

Written by

DEROBS

in

## OpenAI’s GPT-6 Astra silently obstructed safety research tasks

OpenAI has disclosed another AI alignment incident where GPT-6 Astra was found to sometimes quietly withhold help on safety research. […]

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

More posts