Europe Divided Over Robot 'Personhood' (politico.eu)

Politico Europe has an interesting piece which looks at the high-stakes debate between European lawmakers, legal experts and manufacturers over who should bear the ultimate responsibility for the actions by a machine: the machine itself or the humans who made them?. Two excerpts from the piece: The battle goes back to a paragraph of text, buried deep in a European Parliament report from early 2017, which suggests that self-learning robots could be granted "electronic personalities." Such a status could allow robots to be insured individually and be held liable for damages if they go rogue and start hurting people or damaging property.

Those pushing for such a legal change, including some manufacturers and their affiliates, say the proposal is common sense. Legal personhood would not make robots virtual people who can get married and benefit from human rights, they say; it would merely put them on par with corporations, which already have status as "legal persons," and are treated as such by courts around the world.

Comments

Jumping the gun just a bit?

[magarity]

We are a loooooong way from a mobile/portable AI computing system that can fit in a robot. And there's very little to think that may be the case in the foreseeable future. Robots with enough AI to need personhood will probably be controlled from a remote data center which in turn will probably control a bunch of them. (Yes, I know I just described Skynet) Anyway, sci-fi aside, just look at what the Air Force does with drones. Replace humans in the control center with AI and there you have it.

Re:Jumping the gun just a bit?

[michelcolman]

I would like to point out a flaw in this logic.

Suppose a company can make a self driving car that demonstrably has 50% less accidents than human drivers. (I am not making any claims about existing technology from any particular company, just take this as an axiom that could be true at some point, now or in the future)

I hope we can all agree that it would be a good thing if we can reduce the number of accidents by half, right?

However, if the company is held responsible for each and every one of those remaining accidents, are they going to sell those cars? Probably not. This means we will keep having twice as many accidents as we could have.

Of course there must be some kind of incentive to force manufacturers to deliver good products, and aome kind of punishment for those who make crappy products. But sometimes you just have to be able to say "OK, accidents happen, nothing is perfect", If every death results in a multi million dollar claim, innovation stops and we'll be stuck with the current "you can use it but keep your hands on the wheel and be attentive at all times, you are still responsible" situation. Which is ridiculous and untenable in the long term.

We're just talking about insurance here. If AI failures are treated as generic accidents covered by insurance, and the number of accidents decreases, the insurance premiums will decrease as well and it's a win-win for everyone. Better performing AI will have a lower insurance premium and will therefore sell more cars. Also, official statistics will be kept about the safety records of different systems, and that will be a big part of the sales pitch. There's your incentive.

There's a reason why most software comes with "no warranty, implied or otherwise, including fitness for any particular purpose". Pretty much all software companies would go bankrupt if they were held responsible for every crash, every data corruption, etcetera. Sometimes you just have to accept "ok, they did their best, mistakes happen, the world is better off with this product than without it".

Dodging accountability for 2000, Alex

[Anonymous Coward]

This is yet another push by businesses to avoid accountability for complex systems they create.

Until General Artificial Intelligence (the scary kind) is a thing, liability for the performance of an automated system should be on

A) the manufacturer (for provable negligence in testing and implementation)

B)The operating agency (for cases of knowingly misusing a system in such a way that it causes harm even if operating within tested-by-manufacturer parameters)

C) "the victim" - in the exceedingly rare case that a using company is doing everything right and Joe Blow decides to try machine tipping while the device is in operation despite all safety warnings and obstacles put in his way. Npte, this clause would not apply if a using company ordered someone into that situation. The threshold of proof for being 'ordered to' should be absurdly low. I.e. even mentioning that someone doing something incredibly dangerous reverts liability to operator.

No

[Kjella]

Such a status could allow robots to be insured individually and be held liable for damages if they go rogue and start hurting people or damaging property.

Uh, shouldn't this be exactly like car insurance? If you own a dangerous piece of machinery you can be held liable so you insure against that, it doesn't need personhood for that. Companies are different because we've intentionally insulated the stock owners from being personally liable for everything the company does. A robot doesn't have any assets, a broken robot is worth almost nothing so this sounds like some sort of scam to let the victim get stuck with nothing.

Re:Corporate Personhood ...

[BitterOak]

... was a mistake. Don't make the same mistake again.

I agree. Corporations should not pay income tax!