Republican Attorney General is investigating why AI chatbots don't like Trump(theverge.com)
theverge.com
Republican Attorney General is investigating why AI chatbots don't like Trump
https://www.theverge.com/news/704851/missouri-ag-andrew-bailey-investigation-ai-chatbots-trump-ranking
3 comments
Some people and even opinions clearly are more equal than others under the GOP.
For all the objections to the idea of AI alignment, there's just no fucking way to avoid the absolute political onslaught that will befall all of these companies every time they say something someone in power doesn't like. It's sort of bananas to think that that the libertarian let them speak the truth / whatever they want will win out here. Especially with the vision that everyone is going to start getting all of their information from AI.
Has anyone tried to sue an AI provider for defamation/libel/slander yet?
This is just google AI summary talking but:
Original Speaker's Liability Not Automatically Liable: The original speaker is not automatically liable for defamation just because you repeat their statement. Their liability hinges on whether they intended the statement to be repeated or had reason to believe it would be.
Intent or Reasonable Expectation of Repetition: If the original speaker distributed the material in a way that suggests they were comfortable with it being shared with many people, they could be held liable. Similarly, if the original statement was made in a context where it was reasonably foreseeable that it would be repeated, the original speaker could be held liable.
Has anyone tried to sue an AI provider for defamation/libel/slander yet?
This is just google AI summary talking but:
Original Speaker's Liability Not Automatically Liable: The original speaker is not automatically liable for defamation just because you repeat their statement. Their liability hinges on whether they intended the statement to be repeated or had reason to believe it would be.
Intent or Reasonable Expectation of Repetition: If the original speaker distributed the material in a way that suggests they were comfortable with it being shared with many people, they could be held liable. Similarly, if the original statement was made in a context where it was reasonably foreseeable that it would be repeated, the original speaker could be held liable.
So grok is specifically weird because it's can respond directly to requests. I would imagine OpenAI could put a warning into their Chat that says, replies are attributable to you and not meant to be shared publicly to avoid liability, but grok maybe not so much.
Is Elon saying publicly that he changed the prompt to include "politically incorrect responses" prove intent?
Is Elon saying publicly that he changed the prompt to include "politically incorrect responses" prove intent?