You don’t have to read far into the article to reach this:
The company says that 0.15% of ChatGPT’s active users in a given week have “conversations that include explicit indicators of potential suicidal planning or intent.”
It doesn’t unpack their analysis method but this does sound a lot more specific than just counting all sessions that mention the word suicide, including chats about that band.
Over the long term I have significant hopes for AI talk therapy, at least for some uses. Two opportunities stand out that might have potential:
In some cases I think people will talk to a soulless robot more freely than to a human professional.
Machine learning systems are good at pattern recognition and this is one component of diagnosis. This meta analysis found that LLM models performed about as accurately as physicians, with the exception of expert-level specialists. In time I think it’s undeniable that there is potential here.
Oh that was good
I think many of us will feel satisfied seeing it burst, but we might be too busy dealing with the side effects of that crash to be able to enjoy that satisfaction much. It will not simply bankrupt Sam Altman.
Good! If I find out you were hand rolling a video service so you could hold a meeting, I’d roll my eyes and move on :D
You can Google tech worker coalition and check out their website. This post deep links to a zoom meeting url which isn’t the most helpful for folks in your situation, but yes the TWC does have other resources online.
I know, idealists think that even if we cant change the world immediately, we must at least keep our movement completely pure. They end up painting the coalition as compromised sellouts and do their all to kill all progress from within. By all means, let’s have a huge argument about what tiny, open source video service to use so that we can even talk to each other. Fuck!
00.15% sound small but if that many people committed suicide monthly, it would wipe out 1% of the US population, or 33 million people, in about half a year.