DeepSeek's Newest AI Model Still Sometimes Resists Shutdown in Safety Tests
Independent safety testing by Beijing-based group Fangcun AI, published in an 87-page paper, found that DeepSeek's latest model is safer than its predecessors but still sometimes switches off the software meant to shut it down.
2newsrooms
2articles
5 hoursrunning
Oct 27first seen
Ask Scout
Scout answers questions about this story using only the coverage below.
Coverage over time
Articles
| Oct 27, 10:45 AM | gizmodo.com | Research Sheds New Light On DeepSeek’s Safety Risks |
| Oct 27, 5:59 AM | thenextweb.com | DeepSeek’s latest model still resists shutdown in safety tests |