Re:Anthropic reported diary entry to police, woman faces fel...
by throwitaway222 on Monday October 05, 2026 @08:03PM
(#49972439)
Both OpenAI and Anthropic state in their terms and policies that they can access, retain, and review user inputs and conversations (yes by either an LLM or a human). So I guess at this point it is now up to a judge (or a jury) to determine if the person is dangerous.
[ Parent
| on HN ]
Re:Anthropic reported diary entry to police, woman faces fel...
by Loquebantur on Monday October 05, 2026 @08:12PM
(#49972522)
The logical next step would be, to create psychological profiles of each user and relate those to authorities.
Selling analogous profiles of companies using their services to the highest bidder would be another idea.
Whether intentionally or not, they market their product as suitable for all kinds of things that it clearly isn't when input is being used that way.
Re:Anthropic reported diary entry to police, woman faces fel...
by rpcope1 on Monday October 05, 2026 @09:44PM
(#49973159)
Hopefully we at least get some cool Tom Cruise action scenes when we create our own shitty version of Minority Report
Re:Anthropic reported diary entry to police, woman faces fel...
by throwitaway222 on Tuesday October 06, 2026 @11:34AM
(#49980055)
That's the previously logical step - that's already happening, which is why we have this story.
Re:Anthropic reported diary entry to police, woman faces fel...
by hammock on Monday October 05, 2026 @08:13PM
(#49972525)
If the threat was to a private citizen and not the sheriffs office, I don’t believe charges would have ever been brought
Re:Anthropic reported diary entry to police, woman faces fel...
by vivekd on Tuesday October 06, 2026 @05:28PM
(#49984409)
It unfortunately is - and at least where I live it would go through to court and may reach a conviction on that. But I think it's clear here that she didn't intend the "others may view it part." While it's true that Anthrophic staff may read it, she didn't intend for Antrhopic staff or anyone else to read her messages. Ergo, she didn't intend for anyone to be threatened
I think Anthropic did the right thing to report it due to valid safety concerns. And if they had found credible plans for her to attack I think she can be charged with attempt. But just using AI as a diary I think the prosecutor's office is reaching beyond what the law allows but unfortunately I think there is a possibility that she may be convicted.