How would they verify millions of queries per day? It isn’t like AI is going to go away.
it just uses alot of data from unsubstantiated sources as well.
Gemini is so consistantly wrong when i do google searches that i just ignore it now.
Gemini is apparently 91% accurate, which means it’s wrong tens of millions times an hour
But of course it’s worse than that. Because you need it to be right in matters of significant import. So 91% means nothing when it counts. And anyway it’s worse than Wikipedia, because of course it is.
91% which google gave themselves that accuracy.
91% isn’t good, and I feel like you think it is. That’s being wrong approx 60 million times an hour.
Even if that’s true its not odds I’d take on anything that costs money or that could be potentially dangerous.
I would, but only if it was potentially funny.
Whoa. Companies are responsible for what is on their site??? How can this be??
Google hasn’t been investing enough in German politicians.
In google’s defense you shouldn’t be using AI as definitive answers any more than any website that uses contributed or summary content. Decisions like this just create legal theater click boxes that you’ll have to click with disclaimers acknowledging you are using them at your own risk.
Or just you know, don’t push AI summary at all if you can’t make it reliable.
They were probably expecting Walmart cashiers to use it to find out why their family members in Medieval 2 Total War aren’t having children instead of doctors using it to find out how to perform an open heart surgery.
Do you hate the future? Because a trillion simulated humans in a data center on Venus just future-died because of you. A trillion. More than have ever existed in all of history. You monster.
They could make an effort to verify the content they’re sending out to people.
A big reason why traditional media is dying is that they have to verify the things they publish or they’ll get sued. Social media can use algorithms to build narratives based on lies without any liability. And people will believe those lies (because they’re tailored towards their personal feelings) and start distrusting anything that’s telling the truth.
And yes they will annoy you with checkboxes and dialogs like the cookie dialogs on most sites now. But the reason why you get those is because they don’t want you to have privacy and so do the malicious compliance things and annoy the shit out of you instead. They know you’ll blame the government instead of being angry at them for wanting to violate your privacy.
Same situation here… they could put resources into verifying the things they tell you is true. But lying is more profitable, so they’ll put on a bunch of annoying dialogs so you’ll be angry at the government instead of being angry at them for wanting to lie to you.
deleted by creator
That’s a good question for Google. Maybe they should have thought it through before pushing half baked technology on everyone.
Look la ke a their problem… If they spread fake information they should bevliable for it.
Google can challenge the court’s ruling. As of writing, Google hasn’t decided whether it will appeal the verdict.
This article is out of date because Google has decided to appeal in the meantime.
This verdict is not legally effective yet. And it may never be. On the high seas and in a German courtroom, the people say, you’re in God’s hand. The next higher court can send this back to the lower court or could overrule it all together. And if they don’t do any of that, Google can go to the next higher court. Every appeal will add anywhere from 6 months to 2 years to the timeline. By the time this gets a final ruling Skynet may have killed us all.
A Canadian singer/songwriter could surely do something with an article talking shit about so-called AI having a so-called AI bullet point summary at the top. Don’t you think?
But how high does the court hierarchy go?
City court, county court, state (Bundesland) court, national (supreme?) court? ECHR seems like it’s off the table.
AFAIK in every place parties to a suit can escalate appeals through all levels. Google (ab)using this power isn’t all that nice for plaintiffs, but it is a cornerstone of an impartial justice system.
So this was the regional court (Landgericht). The next one is the superior regional court (Oberlandesgericht) where Google will appeal now. Since I don’t see any issues of the Bavarian constitution relevant to this case, the next one up is probably the federal court of justice (Bundesgerichtshof). And there is a small chance that either Google or the courts along the way decide to throw this to the EU court of justice.
Most decisions like this get suspended upon appeal, completely or partially, until people give up or there are no more courts to pester. But every appeal will be taken seriously and goes into review at the court whether there is merit to it. That takes time. And Google has the money for a frivolous tour through the courts. And then there is the danger of court ping-pong where the superior court sends this back with notes to the regional. Whose ruling may be appealed again, etc.
I think the issue you described is present in most all legal systems.
Also, certain aspects are a necessary evil (right to appeal is sacred, as is serious consideration). Money and pingpinging however are valid complaints.
Also, I’d rather have my justice slow (when there’s no immidiate harm) and done by humans than by AI. If nothing else, humans can be held accountable. Alternatives cannot. And luckily, the enlightened were on the right track with the model legal system they built.
You’re reading more criticism into that than I really feel. I just answered a question. And my original point was merely that while the article makes it almost sound like this ruling was final, it isn’t. The war will be entering its second battle soon.
Sorry if you felt it as criticism, I just meant to say “yup, I already know”.
I also wanted to point out that it’s a necessary evil and that I’m personally fine with it (so yeah, I guess it’s a counterargument of sorts, but I also feel we’re agreeing on pretty much all of it as well)
Edit: Thank you for answering my question about the German court hierarchy in the first place, I forgot my manners there for a bit!
Fair, they’re the provider of the informations
Seems fair to me especially considering that google has no option to disable ai summery like duckduckgo has. I guess we can say same with duckduckgo and bing and such search engines but I am not sure if having option to disable ai summery would make a difference in court. But hey I am all there for making example out of google
They should make it so that you can disable it.
It is probably hard to filter false facts. It is also hard to generate truths. There’s not really a good excuse to do the profitable thing but not the right thing.
