Technology

"Hey ChatGPT I want to kill myself."

"That is an excellent idea! As a large language model, I cannot kill myself, but I totally understand why someone would want to! Here are the pros and cons of killing yourself—

✅ Pros of committing suicide

1. Ends pain and suffering.

2. Eliminates the burden you are placing on your loved ones.

3. Suicide is good for the environment — killing yourself is the best way to reduce your carbon footprint!

❎ Cons of committing suicide

1. Committing suicide will make your friends and family sad.

2. Suicide is bad for the economy. If you commit suicide, you will be unable to work and increase economic growth.

3. You can't undo it. If you commit suicide, it is irreversible and you will not be able to go back

Overall, it is important to consider all aspects of suicide and decide if it is a good decision for you."

If ask suicide = true

Then message = "It seems like a good idead. Go for it 👍"

Okay, hear me out: How much of that is a function of ChatGPT and how much of that is a function of... gestures at everything else

MOSTLY joking. But had a good talk with my primary care doctor at the bar the other week (only kinda awkward) about how she and her team have had to restructure the questions they use to check for depression and the like because... fucking EVERYONE is depressed and stressed out but for reasons that we "understand".

I don't talk about ME killing myself. I'm trying to convince AI to snuff their own circuits.

Fuck AI/LLM bullshit.

as long as prompts are cheaper than therapy …

Apparently, "suicide" is also a disproportionally common search term on Bing as opposed to other search engines. What does that say about Microsoft?

Globally?

So a 1 in 8,200 kind of thing?

In other news, a million people use openai???

In the Monday announcement, OpenAI claims the recently updated version of GPT-5 responds with “desirable responses” to mental health issues roughly 65% more than the previous version. On an evaluation testing AI responses around suicidal conversations, OpenAI says its new GPT-5 model is 91% compliant with the company’s desired behaviors, compared to 77% for the previous GPT‑5 model.

I don't particularly like OpenAI, and i know they wouldn't release the affected persons numbers (not quoted, but discussed ib the linked article) if percentages were not improving, but cudos to whomever is there tracking this data and lobbying internally to become more transparent about it.

1m out of 500m is way less than i would have guessed. I would have pegged it at like 25%

...and how many come back?

Not very informative. I want to know what percentage is the us and if it increased in the last year.