![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Found this via a John Scalzi post, and I have to agree with Mr. Scalzi that "this is, I will tell you now, a spectacularly bad idea..." Personally, I want genAI as far away from credit cards (or banking or anything else to do with personal finances) as possible, and then tack another few light years onto that as well.
But, really, while giving dumbasses the ability to allow some LLM to accidentally buy a yacht or infinite paperclips or whatever asinine thing is certainly bad, it's far less scary (and, more importantly, potentially rage-inducing) to me than what was mentioned in this comment under Mr. Scalzi's post there. Granted, credit card companies have been using predictive models for such things for decades now, and that's one thing, but it's a whole different ballgame if they're using "generative AI" (i.e. LLMs) for this, because LLMs are grossly unsuited for that sort of thing.
But, really, while giving dumbasses the ability to allow some LLM to accidentally buy a yacht or infinite paperclips or whatever asinine thing is certainly bad, it's far less scary (and, more importantly, potentially rage-inducing) to me than what was mentioned in this comment under Mr. Scalzi's post there. Granted, credit card companies have been using predictive models for such things for decades now, and that's one thing, but it's a whole different ballgame if they're using "generative AI" (i.e. LLMs) for this, because LLMs are grossly unsuited for that sort of thing.
Thoughts
Date: 2026-06-12 06:56 pm (UTC)From:That's my thought. Followed by, anyone who lets AI shop for them deserves whatever they get. There was even an episode of Love Death + Robots, "The Other Large Thing," in which cats allied with robots, who had access to the household budget.