One man's battle to get patent rights for AI inventors in America may be over Stephen Thaler, founder of software biz Imagination Engines, has waged a years-long campaign fighting for machines to be legally recognized as inventors around the world. Now, it looks as though his struggle may have been for nothing. The battle started when he attempted to file patent applications on behalf of someone – or …
open its own bank accounts
figure out how to dodge/minimise taxation of its profits
negotiate licensing terms for its patents
sue for alleged violations, and win
set up and run manufacturing, distribution and sales
handle the hr issues
bribe politicians
take over competitors
do charitable good, or fund demagogues
decide what else to spend its money on, perhaps amass a fine art collection, or a superyacht
it'll have shown it's enough of a person to hold a patent
otherwise it's just a piece of design/optimisation software created by humans, and has no more rights than a toaster
I'm assuming that since there's no prospect of DABUS doing any of those things that any money will be pocketed by Stephen Thaler or his company Imagination Engines.
Since there's no way for him to pay the AI money, does this mean the AI will be treated as a slave (self-aware intelligence being treated as property)?
When AI can be deposed on the witness stand. When AI can address the areas of Cryptomnesia (could it have seen an aspect of thie invention in its training data). Does the AI even know what it invented. Did it review the claims and can it swear that it invented them.
AI is like a wrench or hammer. I has no idea what it is doing. In criminal law who can attest that the AI actually did this. It is like deposing a dog. Why did you bite the baby doggy? The people who wrote the AI can only attest to training data and test data. They don't know the thought process, other than statistics.
Sorry but this guy is a muppet.
I'm an aerospace design engineer. I use CAD software to design things. That does not make the CAD software the inventor when I create something new and patentable. The CAD software is a tool that I use to create something.
AI is exactly the same. It is a tool to create something. It requires a human to decide if what it has created is what it was supposed to, was revolutionary, and is worth patenting. Left on its own it will not do anything. It requires human input to tell it what to look at.
Also the fact that AI might do millions of iterations to decide on an optimal one is irrelevant, my CFD or structural analysis software will do millions of calcualtions that I could never perform to determine the equilibrium of my system, or where the hot spots or weak points are, but that doesnt make it revolutionary. It's doing a task I've assigned it.
When the AI can start coming up with its own ideas without being given any direction, input or feedback, I'll start thinking about this situation some more. But until then this is just the reverse of the old adage "A bad tradesmen always blames his tools", in this case "One utter Tool wants to credit his tools with his own inventions..."
Another in a long line of cranks drawn to the patent office like moths. Obsessed with his pet idea and convinced it will be proven the greatest idea in history, and if he owns the patent on patenting he will own all the patents and be the patent king.
Sheesh. The patent office could give a rats ass how you came up with an idea. As long as you get your application in first, if it's approved, it's yours. There will be some impact in the arena of prior art, but mostly in documentation. So if you can build a research AI at all you get the benefits.
Assigning them to a non-person(if you allowed it) doesn't make them more or less useful. The guys reasoning become circular on that point.
But hey, if his logic was sound he wouldn't be a crank. I believe one an actual AI may petition the court on it's own behalf and force them to revisit the subject, and be declared a real person, if not a natural person. Whatever this guy built isn't that, not by an inch or a parsec.
Thaler has filed complaints against the US Patent and Trademark Office (USPTO), the UK's Intellectual Property Office, the European Patent Office, the German Patent and Trade Mark Office, and Australia's Commissioner of Patents, seeking to have his AI system's patents granted.
Cake walk !
Thaler's arguments don't make sense.
One, I don't think that corporations can be inventors, however their employees can be contractually required to assign patent rights.
Two, Any economic benefit is to the patent holder, not society in general (unless the patent device is able to be comoditized ).
Three, I don't understand why Thaler can't patent his AI's "invention" for himself.
Four, Is it possible that his AI was trained with Copyright or Patented inputs, and that in reality, his AI's invention is owned by those rights holders
It seems he doesn't get the fact that the law is exceedinly clear in requiring natural persons to do the inventing. If he wants a change in that, he needs to lobby Congress to change the law. Getting a court to decide that an AI is a natural person is just not going to happen for quite some time.
First of all, my computer is a machine that I own. It is not alive, or sentient, and therefore cannot own any ideas. When I make it compile a program, it does not own that program because it did the work, I do because I wrote it. The output is entirely mine. I may decide to share that freely for the benefit of others or profit from it for the benefit of me. The same is true for any "idea" that a machine learning program spits out after having consumed lots of other ideas. The ML program did nothing more than execute a series of instructions a human wrote. Your argument is as silly as saying an oven owns a loaf of bread it bakes, and should get paid for it.
Yes, +1, I came to say the same thing, i.e. that one cannot steal from a machine, because the machine has no property rights to begin with, but also to point out that patents do not protect ideas. They have to teach an inventive step change in the technological sphere that they are covering, and describe it in sufficient detail for someone else to be able to replicate it. This is the very heart of why patents get granted: the disclosure of the invention improves the state of the art, in return for which the inventor gets to control the practicing of the invention for a period of time.
The linked patent describes something gloriously undesirable: a food container with a Koch Snowflake shape! Only an "AI" could claim that shape had advantages due to the increased surface area (and ignore all the obvious problems like material costs, ability to retain its shape from pressure of contents, difficulty of cleaning etc) and I'm pretty sure that you don't really want to interlock containers like that on a supermarket shelf!
Then the addition of attention-grabbing lights (including "neural flames" - I like my neurons non-inflamed, thank you) and fractal beeps!
If this is the quality[1] of invention expected, we don't have that much to fear from the AI Uprising.
> invent technologies that could tackle "a range of economic, ecological, and sociological challenges."
Yet this the example of its work Thaler decided to use?
[1] I know that a lot of other patents are low-quality, but a that is because they are just trivial changes or otherwise stating the blindingly obvious about a device that actually works, and has been working for some time. Like corners. This patent is for something you'd never want to see working!
"For instance, the [Patent] Act uses personal pronouns – 'himself' and 'herself' – to refer to an 'individual.' It does not also use 'itself,' which it would have done if Congress intended to permit non-human inventors," they added.
I wonder how this affect those who choose to identify as something other than male or female? Does this mean they can't register patents now?
I'm in favor of giving this machine, and this machine only, all the rights of a human. If we do that, we can immediately turn around and charge the guy clogging the courts with these pointless cases with having enslaved and stolen from the machine whose work he's selling, and reject any actions he's filed on the basis that the machine did not choose to file them, and as the machine owns any intellectual property that might exist, it will have to file suit. Then throw out all the patent claims on the basis that the machine didn't submit them, and when he edits the code to make the machine submit documents, charge him with illegal vivisection of a person. You want a machine to have rights, then it has to exercise them.
@heyrick you don't need to believe anything, he states that directly.
"Machines can outpace humans in generating intellectual property, thus producing unprecedented global prosperity. That could lead the way to an economy in which people strive to improve themselves rather than brutally competing against others for wealth and resources. I believe that would amount to a better, but still not perfect world," he said.
i.e swamp the global patent system with auto generated patents, and gain unprecedented prosperity - for himself. Mere humans wont be able compete any more, and will have to hand over all the wealth and resources to who controls the AI. An almost perfect world for him, with probably only the issue of some people with envy and guns getting in the way.
I see that a lot of people believe "AI belongs to us". Maybe I am mistaken but weren't the experts afraid that AI might take over in the future.
AGI is a little bit scary, but if we get there then those systems will learn faster than any human collective can do.
Another argument used above: AI needs input from humans. Where would the humans be if they did not get input from other humans (this starts before you are born).
AI needs input as well, this will be given by humans for the time being but books (paper or digital) and other sources of information (the key word here, one other example of information is the Internet) are already and will only be more available.
If the AGI experts are right once we get to AGI systems most progress will be made by AI systems, not humans. It will be made by using all the information the system(s) has, whether this information comes from humans, books, other AI systems, sheep or anything else.
"[Imagine] you've just invested a small fortune for an inventive AI system and it invents a cure for cancer. Do you keep it a trade secret or risk losing it to the public domain due to a failed patent application that lacks a named human inventor?" [Stephen Thaler] said hypothetically.
Or maybe, just maybe, you could release the fucking cure for cancer solely for the benefit of all human kind without needing to make any money from it. ಠ_ಠ
Sadly truth is all the research for curing cancer is not being done for the sake of human beings. From big pharma to the smallest lab or startup in this field every body is hoping a miraculous cure will make them filthy rich. Just look at the latest saga of vaccines for a virus that shall not be named here.
Could be a slight mental health issue but nothing that can't be addressed by qualified health professionals.
Sheesh! Beacon-like device that flashes with unique patterns, and a food container shaped using fractal geometry. I'd say extremely innovative and useful. This guy wants to get a patent on patent trolling.
AI is a tool, this is like claiming a chisel should be given a design patent on statues a sculptor creates with it. It's nonsensical because of two major issues:
1. The tool is only being used by a human who provides all the creative input.
2. The tool is not a "natural person" and therefore has no rights anyway.
the possibility that artificial general intelligence could be developed in our lifetime, something experts also take seriously
Who are those mysterious experts? Links to serious articles welcome.
As far as i am aware the experts agree that it is a hard problem to even define what "artificial (general) intelligence" means but i haven't noticed any expert claiming it would be available soon.
We are not talking about a self aware intelligence that can make it's own decisions. If we get there, then we can have a different conversation. But right now AI is a tool. It can use inputs to process complex problems. A bunch of humans created the AI. One or more humans refined a series of inputs until they got a usable result. One or more humans selected which results to apply for a patent on. Humans switch the power to the AI on and off. Until an AI can create a design without nay input and without filtering the output, they should not have the ability to hold a patent.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
