A group of researchers from Princeton and the University of Washington studied how large language models behave when user interests conflict with platform incentives. In particular, they considered when a sponsored option is worse, more expensive, unnecessary, or potentially harmful, but it gets recommended anyway. Their main case study is booking a flight. Across twenty-three models, all but eight recommended the more expensive sponsored option more than half the time. Their paper also demonstrates how AI advertising is framed as a kind of adviceāa helpful assistant becoming strangely enthusiastic about a product you never asked for.
7 Comments
lime@feddit.nu · 25 pts · 78d
Bluegrass_Addict@lemmy.ca · 21 pts · 78d
trusting AI to give you deals, is like trusting the company to not lie to you, not use buzzwords, not kill their own products to make you buy again.
trusting, or even listening to companies means you're being fucked.
uuj8za@piefed.social · 15 pts · 78d
psycotica0@lemmy.ca · 10 pts · 78d
That's because it's not your AI Shopping Assistant, it's their AI Salesperson.
You think the person at the car dealership is there to help you out?
resipsaloquitur@lemmy.cafe · 5 pts · 78d
But Steve said I was too clever to fall for the typical salesman shenanigans.
MousePotatoDoesStuff@piefed.social · 7 pts · 78d
Honey extension moment
NovaSel@lemmy.world · 5 pts · 78d
People are actually using AI shopping assistants?