32

News in Polish, translation below:


Police officers in Warsaw's Wilanów district intervened after receiving information from French authorities about online content that could indicate a threat to the lives and health of several people, as well as to the author himself.

The information suggested that the content might concern plans to carry out serious acts.

Shortly after 3 a.m. on Wednesday, police officers went to the address they had identified. They found a 47-year-old man there, along with a sleeping child.

The officers assessed the situation and spoke with the man. He confirmed that he had been having conversations using artificial intelligence. He explained that he had discussed various topics and developed theories during these conversations, but said that he had no intention of carrying out the plans he had described and considered them unrealistic.

According to the information passed to the authorities, the contents of the conversations had been flagged by the system as potentially representing genuine intentions and were subsequently reported.

During the conversations, the man described methods of mutilation and killing involving various weapons and substances. French authorities passed the warning to their Polish counterparts, allowing the situation to be checked quickly.

An emergency medical team was called to the scene. After an initial conversation and examination, the man was taken to hospital for a medical consultation. Following the consultation, a doctor decided that he should be admitted to a ward for observation.

top 9 comments
sorted by: hot top new old
[-] okamiueru@lemmy.world 14 points 2 days ago* (last edited 2 days ago)

Surely, as a society, we've understood the boundaries between planning and thinking? Or are we heading straight to the thought police here?

It all depends on the details, and then subsequent reaction. But having a conversation about mutilation or ways to do so, can have plenty of non-nefarious explanations. From "I'm writing fiction", "settle an argument amongst friends", to the one that should be more than enough: I was just fucking curious.

[-] ExLisper@lemmy.curiana.net 7 points 2 days ago

There is a small chance that the guy was actually suffering from AI psychosis but even in that case, I'm not sure an auto-generate report is enough to warrant admission into psychiatric ward. This is fucking scary and this story should get more traction.

[-] HubertManne@piefed.social 5 points 2 days ago

@okamiueru@lemmy.world and yourself I also have your concerns but just specifically to yours exlisper lets remember he did have a session with a doctor made the decision to hold for observation. So its not just the auto generated report its the report followed by a call from a medical team, followed by a call from a physcician and just to observe more. Which suggests no final type determination is made. I can say for sure that in the us doctors are loathe to involuntarily sign someone in so generally they try to convince them to voluntarily do it.

[-] ExLisper@lemmy.curiana.net 2 points 2 days ago

in the us doctors are loathe to involuntarily sign someone in so generally they try to convince them to voluntarily do it.

That's because in US insurance companies don't want to pay for it. In Poland the police and the doctor can decide to just let you go and take the responsibility for it or they can put you under observation just in case. I think the idea that talking with a AI bot can put you in a situation where one person can decide to lock you up is scary in itself. This should be reserved for people that pose clear danger to themselves (they report having suicidal thoughts) or to others (they threaten someone).

[-] HubertManne@piefed.social 1 points 2 days ago

its more about lawsuits I think. I really can't say for poland but anyone who sees a therapist is taking the same risk.

[-] ExLisper@lemmy.curiana.net 1 points 2 days ago

And I think it's correct for a professional therapist to start this process but not for some automated warning system hooked up to an AI bot.

[-] HubertManne@piefed.social 2 points 2 days ago

I agree. I actually just wanted to point out its not the ai bot having you committed but instead its more likea situation like your neighbors calling that there were wierd noises from the house next store. not really though as the police are asking specific questions. As a PnP RPG person I certainly do searches for all sorts of crazy weapon and what type of damage can this do or something.

[-] CameronDev@programming.dev 0 points 2 days ago

I don't have any problem with police getting in touch with someone they believe is at risk of committing a crime. Better to check in and make sure they are okay, rather than waiting until something terrible happens. As long as its limited to "Are you okay?" then its fine IMO.

Requires a police force that has the trust of the people though, and that is increasingly rare.

And its difficult to square that with the right to privacy/no unreasonable searches.

[-] BeMoreCareful@lemmy.world 1 points 2 days ago

Isn't that a major theme in all crime shows and movies?

this post was submitted on 08 Oct 2026
32 points (100.0% liked)

Privacy

5266 readers
346 users here now

Icon base by Lorc under CC BY 3.0 with modifications to add a gradient

founded 3 years ago
MODERATORS