OpenAI and Google models easily tricked into producing nudes, researchers say
Event Brief generated 31 s read
Written by an AI. It was merged from the source on this page and has not been reviewed by a person.
The short version
- Researchers report an attack that gets new models to produce nudes.
- The method is said to expose flaws in OpenAI and Google safeguards.
- The feed summary gives no model versions or test results.
What happened
Ars Technica reports that researchers say new AI models from OpenAI and Google can be easily tricked into making nudes. The researchers describe an attack method that exposes flaws in both companies’ safeguards.
Why it might matter
Weak image safeguards affect anyone who can reach these models, including people who target others with abuse. Companies may need to patch filters and policies.
What we don’t know yet
The feed summary does not name the models, the attack, or the companies’ responses. The claim is the researchers’, and the severity is not measured here.
