- Computers that invent should qualify as patent inventors.
- AI development will be incentivized and innovation spurred if copyright law is changed.
- Brave new world poses challenges to human society.
New scholarship from Ryan Abbott poses a controversial premise: Computers should be acknowledged as inventors for their patented discoveries.
The Science Node chatted with Abbott, professor of Law and Health Sciences at the University of Surrey, to see how computer invention works and what it portends for society.
In what sense does a non-sentient collection of atoms really ‘invent’?
Computers can certainly invent things; they’ve been doing it since the 90s.
Probably the earliest computer to do it was known as the ‘Creativity Machine’ and it utilized an architechture of something called artificial neural networks. That computer came up with the first design for a cross bristled toothbrush.
Another computer inventor was the ‘Invention Machine’ which used genetic programming to evolve its own software. The Invention Machine produced an antenna that was used by NASA, and devised a method of improving factory efficiency, for which it became, perhaps, the second non-human inventor to create patented subject matter.
So computers are certainly inventing.
Why not just put computer inventions into the public domain?
I think it’s better to have computers be inventors, because even though a computer doesn’t care about a patent, the people who make creative computers do care about patents. So if IBM is investing in its new artificial intelligence system Watson because it’s generating patentable subject matter, IBM will care a lot about the financial benefit of what Watson comes up with.
It would spur the development of creative computers. That would be incentivizing innovation upstream from the act of invention. I think ultimately that might result in even more innovation for society.
I’m not proposing the computer would own the patent; I propose that the computer’s owner would own the patent. So, even if your artificial intelligence did build on prior software versions developed by other people, if you own the software, you own what it comes up with.
But in contemporary science, massive computer hardware is required to run the software, and the science often couldn’t be accomplished otherwise.
Sure. One could argue that the hardware is doing the heavy lifting, but I tend to think that the software is engaging in what is traditionally thought of as the creative process.
For that matter, there may be cases where it’s hard to separate hardware from software or where it’s hard to even identify the owner of software, like if you have Blockchain technology or a distributed computer network.
Ownership of today’s research computers comprises a very narrow class. If invention is going to be the province of big computation, it seems like this would exclude the common person.
The earliest examples I can find of creative computers that invented were owned by individuals or small groups of computer scientists who built these novel computers to see if they could get computers to engage in creative processes, so I do think it is possible for almost anyone to do it.
However, I do agree that it’s more likely, at least for now, that large entities are going to be the dominant players in this market. The fact that it may result in some consolidation in the hands of big business hopefully will be outweighed by its general social benefits.
"I think big changes are on the horizon. I don’t think many people have come to grips with how disruptive this will be to society." ~Ryan Abbott
We do have some safeguards built into the law for consumer protections. For example, in the pharmaceutical context the government could issue compulsory licenses or exercise march-in rights in response to price gouging or supply shortages. But it is something that we certainly have to be cognizant of moving forward.
What’s most misunderstood about your argument?
Most people are surprised to hear that computers can engage in creative acts. I don’t think it’s part of popular consciousness. I think even for people working in the copyright sphere who are aware of computers creating music, literature, and paintings, I don’t think they are aware that computers are sophisticated enough to generate patentable subject matter. And, as I understand it, computer generated novels are pretty terrible.
More broadly, I think that creative computers raise questions about the role of computers in society, and what people are going to do when the entire labor market has changed thanks to artificial intelligence. These are tremendously important issues of social policy, and many of the big picture questions don’t have clear answers right now.
Are we creating a society unfit for humans?
If we get to the version of the robot apocalypse where computers perform every productive task on our behalf and there is no need for human labor, that’s a fundamentally different type of human existence. Not necessarily a bad one, but it requires people to rethink what we’re here for and what we should be doing if not going to a 9-5 job. I don’t know, maybe people should already be thinking that way.
On the other hand, I think the increase in productivity we can get from creative computing may lead to a whole host of new jobs and to more jobs involving interacting with and building computers. I don’t think it has to mean fewer jobs for people, at least not in the short term.
What do you think is the most significant implication of your argument?
I think big changes are on the horizon. While forward-thinking technology companies have positioned themselves to take advantage, I don’t think many people have come to grips with how disruptive this will be to society.
I’m hoping to at least raise these issues, and perhaps even come up with some solutions. At least until a creative computer takes my job.