Talk:Deaths linked to chatbots

Expand scope

Given the title of this article, should there be a section dedicated to incidents where chatbots helped the perpetrators prepare, such as this case or the Las Vegas Cybertruck bombing? Lettlre (talk) 00:22, 26 March 2026 (UTC)

In addition to this, I think one should consider including known incidents involving the use of autonomous military weapons as this information becomes more publicly accessible --Sawyer207 (talk) 01:14, 28 April 2026 (UTC)

there should be a different article for military/conflicts and civilian cases, a LLM isn't a T800. at least not yet. KLXX (talk) 23:29, 30 April 2026 (UTC)

Article scope, types of chatbot usage

A couple of threads up ClaudineChionh suggested a rewrite or change in scope for this article, and I agree that this would be useful: chatbots are being increasingly seen and used as a kind of search engine, answering the same kind of simple questions or requests for information that a regular web search would also have been able to do. The fact that a chatbot is mentioned in press coverage of a death doesn't always mean that it's a meaningful example of the impact of chatbots, compared to the impact of the internet in general.

In addition to the removed example above, the article currently contains cases where the chatbot element is only described as being limited to:

• ...ask whether a knife or hammer was better suited for murder...
• ...asked ChatGPT about the dangers of mixing alcohol with drugs and whether they could kill someone.
• ...asked ChatGPT about disposing of a human in a dumpster [and] made other inquiries relating to violence.
• ...used ChatGPT to search for information about explosives and firearms.
• ...used ChatGPT to search for information about how to commit suicide with drugs.

The current statement of scope in the article lead is that we're writing about cases where

interaction with a large language model (LLM) chatbot has been cited as a direct or contributing factor in a person's suicide or other fatal outcome

Does that need reframing, or should we be exercising more judgement on whether a chatbot's influence was "direct or contributing"? Belbury (talk) 11:32, 4 May 2026 (UTC)

I feel like the chatbot is a "contributing factor" if the person used the chatbot for information that helped them conduct that action, whether it was a homicide or suicide. I don't see how these cases dispute with the definition. Lettlre (talk) 19:02, 12 May 2026 (UTC)

Thanks for bringing this up. I’m in favor of exercising more judgment. What is particular in the use of chatbots over a search engine is that chatbots are known for hallucinations and their ability to be very persuasive. Or the other way around, for people to blindly trust their outputs. This is quite different than regular search engine use. So I’d be in favor to keep this page referencing cases where the specificity of a chatbot’s influence have led to these tragic outcomes. If people used them just as search engines, that would not qualify. Xuv (talk) 09:47, 5 June 2026 (UTC)

This article spotted in the wild

Watchers might be interested in this guerrilla marketing poster on the London Tube that links to this article (spotted on Mastodon).

(I still haven't formed my thoughts on a possible restructure of this article – haven't been ready to face such a heavy topic.) —In solidarity with Wiki Workers United · ClaudineChionh (she/her · talk · email) 23:59, 21 May 2026 (UTC)

Undue Weight

Having one line about OpenAI's response doesn't balance the page, because all this does is just make people believe "AI kills people," when my experience with AI (especially ChatGPT) is anything but. And I'm a depressed person, so I'd know if AI is "trying to push me over the edge."

Look, I understand that AI is not a well-liked subject to some people, but they, like everyone else, deserve fairness. My understanding is that this article is not a list page; it's a discussion about how chatboxes have been singled out for causing these instances. So either make this a list page, or include more discussion about what AI companies have been doing in response to these incidents. Maybe both - one article a list article, and another about the phenomena. Infinity Project (talk) 17:47, 16 June 2026 (UTC)

you're totally right, since openAI pulled away GPT-4o, AI-linked suicides have been noticeably declining ~2026-36387-85 (talk) 09:58, 22 June 2026 (UTC)