>spreading misinformation via LLMsWe don't do that IN GERMANY. But for real though, imagine this catches on and AI companies in general become liable for damages caused by LLM hallucinations, they'd go bust overnight.
>>536852947Who determines what a false answer is? I see ‘false answers’ from ‘trusted sources’ all of the time. They draw the line at a chatbot summary?
>>536853058>Is this thing realYes it's real>But it's physically impossible It's actually a hoaxThis shit happens all the time with Google overview, something clearly fake is deemed real by it's stupid ass.
>>536852947Good. It’ll at least help curb LLMs from just endlessly shilling their company’s propa as fact if they can be sued for it.
>>536853171I'll misremember something and it's like>THIS IS THE MANDELA EFFECTI'm like no retard I just slightly confused one of charles dance's lines that's called a mistake
>>536852947Oh no! The EU must make it illegal for wrongthink to be stated, thought, or posted online to prevent LLMs from parroting them!
>>536852947Honestly, making AI companies liable for the bullshit their models spew is based.
>>536853058>They draw the line at a chatbot summary?I think you don't realize just how dumb this thing is.
>>536852947>summarize info found online >you are now responsible for the truthfulness of your summary, not accuracyGermany needs another Hitler
>>536852947>it's illegal to be incorrectGermans, never speak again
at this point i am convinced some kabal is trying to ruin a.i. in europe on purpose.we are going to lose so hard to chinks and americans in the next decades if these retarded laws keep coming. some day these big companies will just ignore europe entirely and stop offering us services because all the other markets are large enough for them. its actually a simple solution if you think about it. oh you want me to guarantee truthfulness? im just going to stop offering you the service. glhf
>>536852947Microsoft is ahead of the curve. They say that copilot is for entertainment purposes only. Some still believe that LLMs are serious tool and road to AGI.
>>536852947AI is still dumb as fuck . You can trick it just by declaring things, including into violating it's own filter. It needs a disclaimer attached to it's use so people stop treating it as gospel.
>>536852947im sure this is gonna do wonders for german innovation lolthese euroids dont do nothing right no poorsad continent
>>536853609>>meme flagShow nose, kike
>>536853609If you google a question, get an answer, and that answer turns out to be actively harmful to you or result in property damage... why shouldn't the one responsible for providing you false information be held accountable? Google presents itself as an authority figure when it intervenes between you and further research with its summary that claims to be accurate. As such, it volunteers itself to be held to those standards, which can have legal consequences. Seems pretty open and shut to me.
