Law AI system 'should be recognised as inventor' - Fine line between The Culture and SkyNet

  • 🏰 The Fediverse is up. If you know, you know.
  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account

An artificial intelligence system should be recognised as the inventor of two ideas in patents filed on its behalf, a team of academics says.

The AI has designed interlocking food containers that are easy for robots to grasp and a warning light that flashes in a rhythm that is hard to ignore.

Patents offices insist innovations are attributed to humans - to avoid legal complications that would arise if corporate inventorship were recognised.

The academics say this is "outdated".

And it could see patent offices refusing to assign any intellectual property rights for AI-generated creations.

As a result, two professors from the University of Surrey have teamed up with the Missouri-based inventor of Dabus AI to file patents in the system's name with the relevant authorities in the UK, Europe and US.

'Inventive act'
Dabus was previously best known for creating surreal art thanks to the way "noise" is mixed into its neural networks to help generate unusual ideas.

Unlike some machine-learning systems, Dabus has not been trained to solve particular problems.

Instead, it seeks to devise and develop new ideas - "what is traditionally considered the mental part of the inventive act", according to creator Stephen Thaler

The first patent describes a food container that uses fractal designs to create pits and bulges in its sides. One benefit is that several containers can be fitted together more tightly to help them be transported safely. Another is that it should be easier for robotic arms to pick them up and grip them.

Container shape
Image copyrightRYAN ABBOTTImage captionThis diagram shows how a container's shape could be based on fractals
The second describes a lamp designed to flicker in a rhythm mimicking patterns of neural activity that accompany the formation of ideas, making it more difficult to ignore.

Law professor Ryan Abbott told BBC News: "These days, you commonly have AIs writing books and taking pictures - but if you don't have a traditional author, you cannot get copyright protection in the US.

"So with patents, a patent office might say, 'If you don't have someone who traditionally meets human-inventorship criteria, there is nothing you can get a patent on.'

"In which case, if AI is going to be how we're inventing things in the future, the whole intellectual property system will fail to work."

Instead, he suggested, an AI should be recognised as being the inventor and whoever the AI belonged to should be the patent's owner, unless they sold it on.

However, Prof Abbott acknowledged lawmakers might need to get involved to settle the matter and that it could take until the mid-2020s to resolve the issue.

A spokeswoman for the European Patent Office indicated that it would be a complex matter.

"It is a global consensus that an inventor can only be a person who makes a contribution to the invention's conception in the form of devising an idea or a plan in the mind," she explained.

"The current state of technological development suggests that, for the foreseeable future, AI is... a tool used by a human inventor.

"Any change... [would] have implications reaching far beyond patent law, ie to authors' rights under copyright laws, civil liability and data protection.

"The EPO is, of course, aware of discussions in interested circles and the wider public about whether AI could qualify as inventor."

Robots building shit is beginning...
 
lol, so when an artist use a pen to draw something, whoever made the pen should now be considered the legal owner of the art, very smarts, thanks for the idea United Kingdumb. Sorry Chris, Sonichu is now owned by Crayola and this site is now probably owned by whatever company made Nulls keyboard.
 
Yeah, much like private centralized blockchains, it's a load of marketing bull intended to attract venture capital. The bad news is VC money might prove to be enough of an incentive to change the law, and eventually we will be living under the yoke of "unbiased and totally fair" algorithms whose "emergent" inexplicable decisions will have to be accepted as the ultimate truth.
 
Inventor and applicant are different things. The important thing is who is the applicant and how did they derive the right to apply for the patent. The AI being your AI would seem like a pretty simple way to solve this, and isn't really much different to employers having the right to apply for patents for their employees inventions.

Great non-issue though academics. Be sure to talk about this a whole bunch and not really do anything useful.
 
Both those patents should be fucking rejected out of hand.

The food container based on fractals? Yeah, that makes no fucking sense. You want smooth surfaces on food storage, otherwise shit gets stuck in the cracks. You also don't want things that stack together to stick when your robot is picking them up, you always want to pick the same amount.

The fucking flashing light pattern that's hard to miss? Yeah, they fucking call that a flashing light pattern. What, did this genius AI try 25 hz on, 5 hz off, 25 on, 5 off?

Also, for existing ai, it needs some criteria to judge whether it has 'succeeded'. Some measurable thing that can be used to determine if it's solving the problem. Did some dude have to watch flashing lights for hours and be like "Yeah, that one was 2% more noticable"?

Sounds like a bunch of bullshit to me.
 
Also, for existing ai, it needs some criteria to judge whether it has 'succeeded'.

What does humanity currently have that qualifies as AI? Unless I'm far behind the times on this topic (quite likely), I thought the gold standard for AI to be considered legit was still the Turing test?

If any of you Kiwis keep up on the subject, by all means enlighten me.
 
What does humanity currently have that qualifies as AI? Unless I'm far behind the times on this topic (quite likely), I thought the gold standard for AI to be considered legit was still the Turing test?

If any of you Kiwis keep up on the subject, by all means enlighten me.
Oh I mean what they call AI, really machine learning. It's not useless, but like you say, it's not really intelligence either.
 
If it doesn't learn in some way, it's not really an AI. And even if it is an AI, even really stupid things can learn in very specific contexts. We're still a loooong way off from Terminator.

The T-800 was equipped with a neural net learning microprocessor (i.e it can modify its own algorithms) and a 120 year nuclear battery. No shit we're a long way off. Interestingly skynet took the learning capabilities out of the T-1000's and later iterations when some T-800 units achieved sentience and started acting in their own interests.
 
What does humanity currently have that qualifies as AI? Unless I'm far behind the times on this topic (quite likely), I thought the gold standard for AI to be considered legit was still the Turing test?
The Turing test is pretty shitty. Even ELIZA managed to fool some people and pass the test.

We're much farther away than people claim.
We still don't know even how general intelligence really works.
Sure you can make an AI that can beat a grandmaster in chess but that same AI would be useless at every other game. They're all highly specialised in a single application.

(Useful) AIs work very differently from an actual brain.
And attempts to fully simulate a biological neural net fail at even simulating the simplest of brains in real-time.

That doesn't stop AI researchers from claiming that human-level AI is just ten years away of course.
They claimed that since the 60s.
 
The T-800 was equipped with a neural net learning microprocessor (i.e it can modify its own algorithms) and a 120 year nuclear battery. No shit we're a long way off. Interestingly skynet took the learning capabilities out of the T-1000's and later iterations when some T-800 units achieved sentience and started acting in their own interests.
Terminator was a fictional movie as well, something worth noting.
 
and isn't really much different to employers having the right to apply for patents for their employees inventions.
Which has historically been abused pretty hard.

The T-800 was equipped with a neural net learning microprocessor (i.e it can modify its own algorithms)
Those already exist. They're still pretty damn dumb. Intelligence in and of itself is a pretty broad category. Honey bee workers learn and imprint on specific flowers, but they're still quite dumb. There's more to anything we'd call a human like intelligence than this.
 
Last edited:
Like sapience in general.
Problem is that while (almost) everyone agrees this exists it's almost impossible to define. Which may be an inherent problem. It's like trying to imagine a state of nonexistence. Your entire experience is based on experiencing things. How can a human conceive of a state of inhumanity?
 
Damn dude, is that how you watch all sci fi? Are you terrified of the day event horizon comes true and we accidentally punch a portal into hell?
That's just stupid. That sort of narrative isn't sci fi its fantasy/horror. While event horizon was a good horror movie, I still hate it for being a typical bait and switch. Similar with Sunshine.
 
Back
Top Bottom