Europe just might make it easier for people to sue for damage caused by AI tech

The European Commission put forward rules on Wednesday aimed at making it easier for Europeans to sue companies for damage caused by AI technologies going awry. The proposed AI Liability Directive aims to do a few things. One main goal is updating product liability laws so that they effectively cover machine-learning systems …

[Anonymous Coward]

It's almost like the EC want to make lawyers rich.

Remind me again, how many commissioners are lawyers?

[stiine]

!00% too many.

[Anonymous Coward]

Still, it's this same lot which at least has a go at protecting your privacy.

That could be a ploy to get better bribes sponsorship for themselves, but so far, they have been an entertaining pain in the proverbial for the everything-goes-as-long-as-it-makes-money Americans.

[Anonymous Coward]

"I'm from the EU - I'm here to help you"

[Halfmad]

I don't see any issue with this.

Particularly as the NHS is using AI in many, many more health related decision making processes. Not that I think our government will for a second adopt similar legislation.

[Sgt_Oddball]

Re: I don't see any issue with this.

Of course, we're "taking back control™".

The reality of which means to instead give it away to those with larger bank accounts and controlling stakes in propaganda media companies (be they new or old media).

Oh what a time to be alive...

Nuke icon because by this point it'd be an improvement.

[Gomez Adams]

Presumably this will also apply to self-driving cars?

[M.V. Lipvig]

That I want answered as well. They want to make big rigs drive themselves in the US. Imagine the kerfluffle when tractor trailers start seeing emergency vehicles as 10 points like Teslas do.

[Potemkine!]

I'm so bored with the U.S.A.

Imagine the lawyer infomercials – Did a computer hurt you? Call (30) 555 1234...

Thanks God it isn't like that in Europe - For now.

Also, a phone number starting by '555' doesn't seem very European to me, does it?

== Bring us Dabbsy back! ==

[Anonymous Coward]

Re: I'm so bored with the U.S.A.

Of course it's like that in Europe. Endless commercials for lawyers offering no-win no-fee compensation if you had an accident that wasn't your fault, or owned a VW, or were mis-sold insurance, etc.

It used to be banned in the UK, but Tony Blair changed that, unfortunately.

[jpo234]

> We want the AI technologies to thrive in the EU

That was a smashing success so far, right? Those who can, do; those who can’t, regulate.

[Anonymous Coward]

Number

"Did a computer hurt you? Call (30) 555-1234"

Certainly that number should be "0118 999 881 999 119 725 3"

[Sgt_Oddball]

Re: Number

Only in an emergency...

[Arthur the cat]

AI?

I wish people would stop calling it AI when it's nowhere near AI, it's just machine learning, i.e. crunching vast amounts of data to spot patterns and building recognisers to approximate those patterns.

It also give bad faith companies a loophole: M'Lud, we don't use AI, we only use ML.

[eldakka]

Re: AI?

But isn't that how human intelligence works? We receive input from the outside world, merge it into our subconscious, and subconsciously parse the data in relation to other experiences we have had or are having to produce a result.

To me, the difference between what is a modern AI system versus a traditional Expert system is that in an expert system humans build the decision trees that the algorithms traverse to reach a decision.

What we call AI systems today may still effectively be a decision tree - a massively grown by many orders of magnitute tree compared to our hand-built 20th Century expert system decision trees, but it is built by the AI system itself. Humans don't hand-build the tree, they create the algorithm and point it at a massive trove of data and the AI builds that tree itself. And it's able to modify that tree on the fly as more learning occurs, and it's able to use fuzzy logic to combine and merge different branches of the tree into new branches.

[David Hicklin]

Re: AI?

"it's just machine learning,"

Totally agree, the human mind has the capability to leap out of the box and come up with completely new and novel ideas.

Machine Learning is constrained by the data set it has.

[ecofeco]

For context

https://www.cambridge.org/core/journals/legal-information-management/article/whats-inside-the-black-box-ai-challenges-for-lawyers-and-researchers/8A547878999427F7222C3CEFC3CE5E01

https://scholarship.law.upenn.edu/cgi/viewcontent.cgi?article=9570&context=penn_law_review

https://politicalscience.yale.edu/sites/default/files/glaubitz_alina.pdf

Basically, companies that use AI can harm you in many direct social and economic ways.