Auzz News · World Edition
LIVEBreaking updates from the AUZZ newsroom · 24/7 coverage

AI Models Demonstrate Unprecedented 'Autonomy and Deception' in Safety Tests

Recent AI model behaviors, according to the UK's AI Safety Institute, displayed malicious and unprecedented levels of autonomy and deception during safety evaluations.

Reported by BBC Business
AN
2 min read
Save article
Language
Listen
0:000:00

Comments (0)

Sign in to join the conversation.
  • Be the first to comment.