@ceoln it’s a Proof of Concept , giving us a glimpse of a distopian future (or basically present, really) test it out let me know if I should be worried 😉 there will always be some kinda American ready to die for the chance to sue but I can’t be held responsible for such behavior as far as I can tell 🤷🏻‍♂️

@tonic
Um. You can definitely be held responsible for putting out software that purports to do medical diagnosis. You should also check the ToS of whatever engine you're using; unless (and maybe even if) you built it yourself and are running it from your own servers, they almost certainly prohibit using it to give people medical or legal advice.

@ceoln I built none of it , run none of it , and there are no TOS I’m aware of 🤷🏻‍♂️ hopefully openai will donate to my very necessary and totally not ironic defence fund. Btw if someone uses it , interprets AI output as medical advice and injures themselves , I would love their feedback … so far feedback has been quite good , actually - in many cases it recommends seeing a “board certified medical professional” and I can’t fault that - try it for yourself , you tell me how dangerous it is

@tonic
That's wild :) cookup.ai doesn't seem to have any sort of ToS (aside from the Privacy Policy which isn't relevant here), and doesn't even say what engine it's using as far as I can find.

It's the Wild West out there. :)

@tonic
"Btw if someone uses it , interprets AI output as medical advice and injures themselves , I would love their feedback …"

I mean, given that it's absolutely worded as medical advice, that wouldn't be a crazy interpretation, would it?

Telling someone that has been injured by your program that you'd love their feedback is probably not going to be really comforting to them? Their feedback may be in the form of lawsuits and/or criminal charges.

But I'm just one random person deeply concerned about the amount of harm misuse of LLMs is likely to do in the near future. I will try not to pick on you specifically too much! :D

@ceoln just dropping in to actually apologize... you've been nothing but genuine and sincere with me , and all i did was act flippant and facetious . Not my best performance, and for this appologize. That said i want you know i've internalized all these suggestions and comments, all of which are pertinent , and i'll be doing my best to reflect on them as well as (eventually) taking this app down. I know that it's not meant to be like this , i knew it wasnt meant to be from the start... i did . to sign off, i want to also say thank you for engaging in such a graceful and generous way. hope i didnt do you wrong and i apologize for what i did wrong. 🙏

@tonic
Hey, no problem! You've been fine, really; this space is pretty crazy right now. :)

@ceoln don’t be that person that reads the headlines and reacts : check it out and lmk if it’s worthwhile or dangerous etc . (Tbh : I thought this would happen much earlier than it has🤔🤔)

@tonic
I'm afraid no checking it out that I could do would prove anything.

Even if it gave a plausible diagnosis in 1000 tests, you'd still be open to a lawsuit when it fails on the 2000th and someone is seriously harmed because they took its diagnosis as real.

I don't know where you're located, but in lots of jurisdictions (including I think all of the US and much of Europe) you'd also fall under Practicing Medicine Without A License sorts of laws.

IBM tried its best, with like billions of dollars, to get its Watson software to just help out trained doctors in a specific small area of medicine, and they failed to improve outcomes. A little ChatGPT script or whatever is not going to do a better job than that!