Scivillage.com Casual Discussion Science Forum

When AI can create anything, what makes human creativity matter?
https://www.eurekalert.org/news-releases/1148455

INTRO: As artificial intelligence becomes capable of producing sophisticated art in seconds, a new paper challenges us to rethink what makes human creativity valuable. The paper argues that while AI can bypass years of artistic training, it cannot replace the transformative experiences that give creativity its deeper meaning. The research suggests that the future of art may lie less in producing impressive works and more in using technology to inspire exploration, meaningful experiences, and new ways of seeing the world. (MORE - no ads)

---

Does AI know when to say "no"? New study reveals AI chatbots recognize unnecessary medical tests, but stumble where human doctors disagree
https://www.eurekalert.org/news-releases/1148434

EXCERPTS: In an era where patients and clinicians increasingly turn to artificial intelligence for medical advice, a critical question emerges: does AI know when not to test or treat? A new joint study from Ben-Gurion University of the Negev (BGU) and Clalit Health Services reveals that leading Large Language Models (LLMs) successfully identify when medical interventions are redundant in the vast majority of cases, stumbling primarily in gray areas where medical consensus itself is divided.

[...] A human physician must stand firm before an anxious patient, whereas language models tend to go along with the person in front of them. The next critical research question is whether an AI assistant will maintain the correct clinical restraint when an anxious or insistent patient pushes back." (MORE - no ads)
_

LLMs have no will with which to push back against a human and are designed to be accommodating. Trusting them with any kind of judgement call is irresponsible.

Syne Wrote:LLMs have no will with which to push back against a human and are designed to be accommodating.

That's so sweet. See a lap-dog .. love the lap-dog .. all dogs are lap-dogs. Actually not. AI red in tooth and claw just does what it's supposed to do. In this case (probably) .. if an MRI scan costs $1,000 and you don't give 1,000 unnecessary scans .. you've saved $1 million. If one unscanned person successfully sues for $1 million you might as well have scanned the lot. The signs aren't clinical .. which is where AIs are useful.

But doctors, who do currently avoid unnecessary medical test, are subject to malpractice lawsuits. I'm fairly certain that all AI advice has warnings and addenda to absolve it of any liability. That's how AI companies keep from being sued for all the chatbots that encourage suicide, helping plan a murder, etc..

(Today 01:28 AM)Syne Wrote: [ -> ]But doctors, who do currently avoid unnecessary medical test, are subject to malpractice lawsuits. I'm fairly certain that all AI advice has warnings and addenda to absolve it of any liability. That's how AI companies keep from being sued for all the chatbots that encourage suicide, helping plan a murder, etc..

So you think insurance companies having AIs sitting in on consultations are wasting their money. A valid opinion and I am happy to give it the respect it deserves.

I seriously doubt insurance companies are trusting their business to AIs. I'm betting there's a human who makes the final call... after the AI summarizes the actuarial tables.