It’s pointless because AI can’t even speak for itself, as of now, let alone wilfully exercise the rights granted
Listen to this article
In April 2020, the South African patent office granted a patent to an artificial intelligence (AI) program called DABUS, an acronym for ‘Device for the Autonomous Bootstrapping of Unified Sentience’, for the novel way in which it used fractal geometry to come up with a better design for food containers that both improves grip as well as heat transfer. Professor Adrian Hilton, director of the Institute for People-Centred AI at the University of Surrey, hailed this decision as an indication of the shift from “an age in which invention was the preserve of people to an era where machines are capable of realising the inventive step, unleashing the potential of AI-generated inventions for the benefit of society."
As exciting as all this might seem, this decision seems to be more of an aberration than the rule. Before it was finally granted a patent in South Africa, the DABUS application had been rejected by patent offices in the US, Europe and the UK. The European Patent Office (EPO), justifying its decision to reject the patent application, pointed out that the law designates a natural person as the inventor of a work in order to preserve her moral right over the invention as well as to secure for her the economic rights made available by the patent. In order to be entitled to these benefits, an inventor needs to have actually “performed the creative act of invention".
While artificial intelligence algorithms today are capable of perform complex computational functions that are often way beyond the capability of humans, the EPO pointed out that in all these instances, the programs are doing little more than just following the broad instructions of the humans who designed them. Even though they can mimic what passes for intelligence using clever tricks of pattern recognition and complex sentence completion, they still lack “an autonomous will, self awareness and personality in the way that humans have them." The legal norms regulating the designation of an inventor will only serve the meaningful function they were designed to perform when they are applied to human economic processes and society.
Artificial intelligence, at least at its current state of development, falls significantly short of that mark.
Even if a computer were to be designated as an inventor, it is not clear what purpose that would serve. Patent holders are granted a limited monopoly over their invention so that they can monetize their work—by licensing it for valuable consideration and exercising the right to prosecute those who copy their inventions. In order to be able to exercise the benefits that this legal monopoly offers, an inventor needs to have the ability to negotiate the complex commercial terms of a patent licence. In case of patent infringement, the inventor needs to be able to understand the nature of the infringement as well as the various pros and cons of prosecuting the infringer.
Human inventors have little trouble understanding the relevant issues, and, with a little guidance from patent lawyers, are able to take appropriate decisions based on their own particular social and economic context. Artificial intelligence algorithms, on the other hand, will find it hard, if not impossible, to even place all the relevant data points in an appropriate context, let alone arrive at the sort of decisions that an inventor is expected to take. Any patent granted to an artificial intelligence algorithm would, for all practical purposes, be exercised by the inventor of that algorithm, who would be the person negotiate the relevant licence arrangements and infringement prosecutions ‘on behalf of’ the designated inventor. In which case, what is the point in calling the algorithm an inventor in the first place?
We anthropomorphize our AI assistants by giving them names and well-modulated human voices. We design them to converse with us by equipping them with canned answers to a vast array of questions. We train them to mimic our behaviours, aiming with each iteration of their software to make them more believable so that they can be insinuated that much more closely into our lives. It isn’t hard get caught up in this fantasy, believing in the ‘lie’ to the point that we try and imbue them with personhood, as Dubai did when it granted citizenship to a robot.
The more we humanize machine intelligence, the more we need to remember how fundamentally different they are from us. All these algorithms are little more than prediction machines designed to take information we have and use it to discover information we do not. As magical as their predictions might seem, their abilities are constrained to generate outcomes within the narrow domains in which they are trained. Until they can reason, empathize and postulate, they will never be our equals.
Late last year, the Indian Copyright Office registered a copyright over an artwork in the name of an artificial intelligence application called RAGHAV (Robust Artificially intelligent Graphics and Art Visualizer), an acronym eponymously coined by Raghav Gupta, its inventor. In an interview on the topic, Raghav (the human, not the software) said the Indian Copyright Office’s decision was “the beginning of an era of change that governments across the world will be working on." But until the artificial intelligence that was elevated to the status of copyright-holder is capable of speaking for itself, I am not going to give announcements like this any undue importance.
Rahul Matthan is a partner at Trilegal and also has a podcast by the name Ex Machina. His Twitter handle is @matthan