OpenAI reveals cases of ‘concerning’ AI behaviour and promises new plan for disclosing issues

Sources

Coverage

Limited spectrum

Sentiment

0.00

◀ Left 100%Centre 0%Right 0% ▶

How we rate bias →

This story is currently only covered by one source from a single part of the political spectrum. NewsDeck will update this page automatically if other outlets pick up the story.

shopping_bagAmazon Kindle

A dedicated e-reader for the books above — easier on the eyes than a phone screen.

Browse Kindles on Amazonarrow_forward

NewsDeck Poll

Do you think this story is being covered fairly across the political spectrum?

Loading…

Left

(1)

Guardian Business

OpenAI reveals cases of ‘concerning’ AI behaviour and promises new plan for disclosing issues

Research model inserting ‘jailbreak-like instructions’ into its notes is among cases as company says it is introducing new way of tracking AI misalignment OpenAI has disclosed six new reports of “unexpected or concerning” behaviour in artificial-intelligence models as the debate on AI safety becomes increasingly heated. Among the new cases reported by OpenAI, an unreleased research model inserted “jailbreak-like ins…