Comment by SilverElfin
5 days ago
Why is it good? Everyone with common sense knows AI can be wrong. And it’s not buried in their TOS. It’s in the chat box. But even if it wasn’t, it’s ridiculous to create liability for AI chatbots.
5 days ago
Why is it good? Everyone with common sense knows AI can be wrong. And it’s not buried in their TOS. It’s in the chat box. But even if it wasn’t, it’s ridiculous to create liability for AI chatbots.
it’s ridiculous to create liability for AI chatbots.
Liability isn't being *created* here, it has existed legally for a very long time.
False information can cause real harm --- and the legal burden of proof is on the source.
Search engines were provided legal exemption on the basis that they were simply quoting/referencing 3rd party sources who where legally liable for the content.
LLM chatbots legally exceed these bounds by fabricating info/content on their own ---data that does not exist elsewhere. This is a liability issue waiting to happen as there is no other responsible party/source to blame.
AI overviews cites it's sources though...
Yes, AI would never list irrelevant sources and fabricate links.
https://blog.platinumids.com/blog/ai-hallucination-crisis-co...
2 replies →
And often still makes claims that completely contradict those sources it's "citing".
Common sense is so rare it might as well be a super power.
e.g. The failed marketing of A&W's third-pounder burger as so many people (in the USA) didn't believe that it was bigger than a quarter-pounder.
https://en.wikipedia.org/wiki/Third-pound_burger