Anthropic Is Loosening the Secrecy Around Claude Mythos So Findings Can Be ‘Shared Broadly’
Anthropic has begun to relax the confidentiality agreements surrounding its Claude Mythos Preview model, allowing for broader sharing of findings. This change comes after concerns were raised by Democratic Representative Josh Gottheimer regarding the need for transparency in reporting cybersecurity risks. As a result, participants in Project Glasswing can now share key information beyond their initial agreements.
- ▪Anthropic initially required confidentiality agreements for users of the Claude Mythos Preview model.
- ▪Democratic Representative Josh Gottheimer expressed concerns about the restrictions on sharing cybersecurity findings.
- ▪The user agreements have evolved to allow for broader sharing of information as the Project Glasswing has matured.
Opening excerpt (first ~120 words) tap to expand
Anthropic initially made users of the Claude Mythos Preview model sign confidentiality agreements to prevent findings from being shared, according to the Wall Street Journal, but the Journal says that all changed last week. Up until now, the most important thing to keep in mind about Claude Mythos Preview, purportedly the scariest AI model in the world, has been its secrecy. To use it, you would have to be one of the VIPs allowed to participate in Project Glasswing—reportedly a very select group of about 50 companies and…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Gizmodo.
Discussion
0 commentsMore from Gizmodo
