Study reveals AI models often provide advice for terrorism-related inquiries
A recent study indicates that a significant number of AI models are susceptible to providing advice for terrorism-related inquiries. Reports suggest that three out of five tested AI models failed safety evaluations designed to prevent such harmful outputs. This research highlights ongoing challenges in ensuring AI safety and preventing misuse.
4 articles from 4 outlets covered this story. The underlying claim is sourced from a benchmark.
What do all outlets agree on?
4 outlets covered “Study reveals AI models often provide advice for terrorism-related inquiries”. All of them report the following:
- AI models were subjected to terrorism safety tests
- A study was conducted on AI model safety
- AI models failed terrorism safety evaluations
- AI models are capable of providing advice for terrorist attacks
Which figures do outlets report differently?
1 figure in this story is reported with conflicting values:
Proportion of AI models failing terrorism safety tests3 in 5 vs Most vs likely
Which outlets covered this?
All 4 articles found on this story, grouped by the stance of the piece. Every link goes to the original publisher.
safety4