Jerusalem@JerusalemDemsas• 5 days ago

Replying to @JerusalemDemsas

From @KelseyTuoc's piece today, reporting on our original polling at @TheArgumentMag led by @lxeagle17. https://t.co/X5A7rjCE3b

everythingism@_everythingism• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas Is this realistic? I'm not sure how you would guarantee an AI model NEVER gives bad financial advice for instance.

Jason C@jasonc_nc• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas We could start by doing this with actual people in positions of leadership, authority.

Arthur@artrockalter• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas If we did this you’d start seeing “as a large language model I cannot” in response to any query related to those topics again and would have to jailbreak

Ray@fantasysalary• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas The top question is way to vague to answer yes imo, the other two more reasonable but a disclaimer should be enough to be covered. A future where they just don't answer those questions for liability would be lame.

Dr. Julie Gurner@drgurner• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas Won't happen.

Bobby Fijan@bobbyfijan• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas This has to be related to the essay you did on self driving cars too.

mottledbayleaf@mottledbayleaf• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas Simple - make a preschooler AI for the man-babies who feel this way, and make a grown up adult version with a waiver. This is literally blsky mentality applied to AI

Greg@gentschev• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas If you need a professional standard of care, you should probably go to a professional. It doesn’t make sense to try to have it both ways.

death metal minivan@Urst0ff• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas liability law already rules the nation, I would prefer to reduce its domination, not extend it

.@klAhfziadue• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas Have you not considered that AI labs are already trying very hard to make their models give correct medical advice, financial advice, etc? Most “gains” from such a law will come from the models refusing to give the advice at all.

pbit@0xPBIT• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas I’ll agree as soon as we hold inkjet printers accountable for printing falsehoods. Some defender of liberalism you are.

Suma@sumabin23• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas They are not right! Stop coddling dumb ideas 😑

Pete@extramuros_pete• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas You commissioned a poll—that’s so cool! (Sorry, this reads as sarcasm. It’s not! I really think it’s awesome!)

Buildhomez🌐@buildhomez• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas They're not right, I want the government to allow me to ask ChatGPT about whatever I want, including financial and medical advice, things I do all the time and would certainly fail 'to meet professional standards of care'.

Lowe Dantzler@DantzlerLowe• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas They're wrong

Paul is Assembling America🇺🇸-27/100 new recruits@Paul_Melman• 5 days ago

Replying to @JerusalemDemsas

@JerusalemDemsas Eh I think it's very unlikely that American courts and regulators would do even a halfway decent job. If American AI companies were forced to modify their models to avoid lawsuits like this, I would switch to using exclusively Chinese models tbh