AI: Article: How dangerous is Anthropic’s Mythos AI?

[Kim Locke]

Just what we need, an AI that can hack.

https://www.theguardian.com/commentisfree/2026/may/08/how-dangerous-is-anthropics-mythos-ai

[Debra Turnbull]

@Kim-Locke

Thanks for sharing Kim. Yes, I heard about Mythos AI having the ability to re-identify de-identified and anonymized (I need to find the resource on that) data.

This has implications in the clinical trials & drug discovery business where the need to protect patient identity is paramount.

[Kim Locke]

@Debra-Turnbull I think this should raise a red flag for PAN, in general, in terms of privacy and consent of patient identity. There's a lot of red flags with AI, but this one is pretty big.