[lanfeust6]

This stat is for AMI, for any mental disorder ranging from mild to severe. Anyone self-reporting a bout of anxiety or mild depression qualifies as a data point for mental illness. For suicide ideation the SMI stat is more representative.

There are 800 million weekly active users on ChatGPT. 1/800 users mentioning suicide is a surprisingly low number, if anything.

[JDEW]

> 1/800 users mentioning suicide…

“conversations that include explicit indicators of potential suicidal planning or intent.”

Sounds like more than just mentioning suicide. Also it’s per week, which is a pretty short time interval.

[robocat]

But they may well be overreporting suicidal ideation...

I was asking a silly question about the toxicity of eating a pellet of Uranium, and ChatGPT responded with "... you don't have to go through this alone. You can find supportive resources here[link]"

My question had nothing to do with suicide, but ChatGPT assumed it did!

[maplethorpe]

I got a suicide warning message on Pinterest by searching for a particular art style.

[btilly]

We don't know how that search was done. For example, "I don't feel my life is worth living." Is that potential suicidal intent?

Also these numbers are small enough that they can easily be driven by small groups interacting with ChatGPT in unexpected ways. For example if the song "Everything I Wanted" by Billie Eilish (2019) went viral in some group, the lyrics could easily show up in a search for suicidal ideation.

That said, I don't find the figure at all surprising. As has been pointed out, an estimated 5.3% of Americans report having struggled with suicidal ideation in the last 12 months. People who struggle with suicidal ideation, don't just go there once - it tends to be a recurring mental loop that hits over and over again for extended periods. So I would expect the percentage who struggled in a given week to be a large multiple of the simplistic 5.3% divided by 52 weeks.

In that light this statistic has to be a severe underestimate of actual prevalence. It says more about how much people open up to ChatGPT, than it does to how many are suicidal.

(Disclaimer. My views are influenced by personal experience. In the last week, my daughter has struggled with suicidal ideation. And has scars on her arm to show how she went to self-harm to try to hold the thoughts at bay. I try to remain neutral and grounded, but this is a topic that I have strong feelings about.)