Comment on Missouri AG: Any AI That Doesn’t Praise Donald Trump Might Be “Consumer Fraud” (No, Really)
M0oP0o@mander.xyz 4 days agoThe ability to make shit up helps as well.
Comment on Missouri AG: Any AI That Doesn’t Praise Donald Trump Might Be “Consumer Fraud” (No, Really)
M0oP0o@mander.xyz 4 days agoThe ability to make shit up helps as well.
Vupware@lemmy.zip 4 days ago
Certainly! that’s why you use them to find sources. I don’t immediately trust anything an LLM spits out.
M0oP0o@mander.xyz 4 days ago
The amount of times I see people just take what is outputted as gospel is scary. And on things with real risk like electrical repair!
Vupware@lemmy.zip 3 days ago
Absolutely. It is essential to use these LLMs with an abundance of caution, and to explore a broader range of sources from various viewpoints.
Caution and breadth of sources should be staples of any historical or contemporary investigation, but alas, the minds of my fellow citizens have been eroded; these practices cannot be assumed to be broadly adopted, and as such, when saying something positive about LLMs, we must provide ample disclaimers and remind the general populace of the dangers and harms these technologies expose us and our surroundings to.
13igTyme@lemmy.world 4 days ago
I’ve seen Google AI say one thing and provided the source. When I read the source, sometimes it’s completely different or even the exact opposite.
M0oP0o@mander.xyz 4 days ago
Oh yeah the LLM that is built to make things up at the drop of a hat can also make up the sources, its wild that people can still trust a literal machine of lies.
I saw one the other day that stated AC and DC power where interchangeable…