Play a game with me: imagine that scientists at MIT have succeeded in creating beings of artificial intelligence (AI). These machines can think and feel just as humans do. They demonstrate the capacity to act with intent, to make ethical decisions, and to learn from experience. They are composed of silicon circuitry or biologically engineered computational molecular arrays derived from an analysis of the brain’s “software”. As inventions, they clearly possess the requisite originality or “non-obviousness” for patentability. But in the regime of AI, a new and yet unanswered question for patentability presses upon us: as free-thinking moral and conscious agents, do they fall outside the conceptual boundaries of “property”?
The above scenario is actually less theoretical than it first appears. Many prominent scientists, including MIT computer science professor and “futurist” Ray Kurzweil, have written books (see for ex. Kurzweil, “The Age of Spiritual Machines”) predicting that humanity will be forced to confront this very question quite possibly within twenty years. Indeed, the Media Lab at MIT has already succeeded in producing quite sophisticated forms of AI called “expert systems”. Some are used to assist physicians in diagnosing diseases and others will be used to create the proposed “smart grid” for saving electrical energy. These individual applications of artificially intelligent systems and the efficiencies they create are of great practical value, with the ability to save lives and money. However, they represent only the first iteration of the evolution of devices into creations which may eventually become creative or even “aware”.
If artificially intelligent devices could be said to create ideas which never existed and interact with their environment in a way that we would recognize as “emotional”, should these devices receive the dignity of a more human appellation, and would they then be eligible for protections under the law? Should they possess rights to “life” and “liberty” – and what would those rights possibly entail? May we treat “consciousness in a box” as intellectual property, enslave its actions and own the fruits of its labor?
Just as patenting genes is seen as perhaps reducing the building blocks of nature to property, the evolution of artificial intelligence will pose yet more difficult issues. If an expert system can explore, understand, quantify and improve upon fundamental processes such as manufacturing, chemical engineering, drug development and so on, the intellectual property will not only be the system but its capacity to produce further invention, ultimately reinventing itself. This begs the question of whether the process of thought itself may be “property” if it can be replicated artificially. May the process of thought then also be enslaved, or will intelligent devices then be entitled to freedom from ownership, and the fruits of their thoughts also for the common benefit of all?
This is a great question to raise, and I think intuitively at least the idea of “owning” an intelligent machine is extremely problematic. On a more structured note, Kant insisted that the basis of ethical rights was rationality and self-consciousness, not any particular biological structure. So under a Kantian ethical/legal framework at least, a truly intelligent and “aware” machine would be entitled to all the rights that we are accustomed to – including the right not to be enslaved.
This raises another question–if we have intellectual property that is capable of having ideas, who owns that intellectual property? What if a patented AI machine comes up with an invention, can it patent it? And if so, does the owner of the patent on the AI machine own the AI’s patent on the invention? Weird things start to happen when you can claim ownership of a thinknig being….
I think the process by which the intelligent machine is created can be claimed as IP, but not necessarily that which the intelligent machine creates.
Assuming original AI, however, is a huge hurdle that we might not want to take for granted so easily (even for the sake of the subsequent IP discussion). A machine is limited to the functions which are coded into it…it comes down to a series of 1s and 0s compiled by a human. Everything that the machine does, it does BECAUSE deliberate human actions have been taken to make it capable of doing so. Basically, the actions of a machine are intrinsically unoriginal (despite what strong AI supporters might say). This might make the discussion of AI’s IPRs (or lack thereof) a moot point…
A human may have created the parameters for the machine, but in the scenario James refers to, the human does not manually input all possible actions available for the machine. The fact that the machine has the capability to create new ideas/new patterns within the parameters that its creator gave it makes this very troubling. It seems almost human-like. If I were theologically-inclined, I might put a plug for the fact that all humans are limited, inherently, by certain parameters/code/abilities. Yet, it hasn’t stopped us from doing many interesting things within our limits and being somewhat creative (arguable, I know).
Frankly, all of this sounds a lot like playing god.
This post reminded me of the work-for-hire doctrine of copyright law. In most cases, intellectual property produced by employees while they are on the job or utilizing resources from their company belongs to the firm. Similarly, if a robot were “employed” by a company and devised a new innovation, it seems logical that the company would be able to claim patents and other intellectual property produced by the robot. Absolute artificial intelligence probably will not be seen for a long time, and most robots act according to human parameters. If a robot were programmed to try mixing thousands of combination of chemicals, for instance, to create a new drug, why shouldn’t the drug company be able to patent a successful drug that results?
The broader idea of companies exploiting machinery is nothing new. During the industrial revolution, Marx was highly critical of the fact that machinery-what he called fossilized labor–was used to both displace human labor and cause the ruling class to profit from the past work of the proletariat for years to come.