Published on Data Blog

Inventors and innovations in the era of AI

This page in:

artificial intelligence, future technology and business concept
Image: Syda Productions/Shutterstock

Since the time of cavemen, every technological discovery has been made by a human being: the digging stick, the scraper, the wheel, the car, penicillin. For many years, nobody challenged a worldwide legal approach that only a human being can be an inventor. It was until an American scientist and inventor, Doctor Stephen Thaler created an artificial intelligence (AI) Device for the Autonomous Bootstrapping of Unified Sentience (DABUS) and tried to change the established view. In 2018-2019, Dr. Thaler conducted a fascinating experiment that was eventually only a first step in a long road to understanding if AI can be a creator, innovator, and inventor.

The lack of certainty here may lead to long-lasting consequences for innovation and the economic development of nations. Observers believe that if AI-devised inventions are unable to be patented, there may be less investment in this technology. In this case, AI-generated products should be placed in a public domain. Alternatively, AI works protection may incentivize investment in AI and the use of AI systems. Though, a proliferation of patents covering AI-devised inventions might have a detrimental impact on R&D, leading to a decrease of researchers and, generally, jolting the current intellectual property regime, which relies on the "human inventor" rather than on “AI inventor”.

DABUS, the inventor

This discussion became hot after DABUS, without any human help, created an improved food container that used fractal geometry to change its shape. Dr. Thaler submitted patent applications in several countries naming AI as an inventor. The prosecution of applications went in different directions in different countries.

The patent offices in the UK, EU, and the US rejected granting inventorship to an AI machine, pointing out that an inventor must be a human being, not a machine. Nevertheless, the patent applications went off the road in South Africa and Australia. In June 2021, South Africa became the first country globally to accept the possibility for an AI to be an inventor.  Now you may find the following text in the inventor’s name line: “DABUS. The invention was automatically generated by an artificial intelligence”.

The following month, the Federal court of Australia sided with Dr. Thaler and overturned the Australian patent office’s decision to reject granting inventorship to an AI machine. Here is a very testimonial paragraph from the court’s opinion: “…Who is the inventor? And if a human is required, who? The programmer? The owner? The operator? The trainer? The person who provided input data? All of the above? None of the above? In my view, in some cases, it may be none of the above. In some cases, the better analysis … is to say that the system itself is the inventor. That would reflect the reality”.

A 2D artwork  created using Artificial Intelligence

2D image generated by DABUS

Threat to human-devised inventions?

Indeed, data-driven AI is rapidly becoming ingrained in many facets of society, from voice and facial recognition systems to automatic language translation services to customer service chatbots and virtual assistants. But as with all great inventions, advances in AI also present new concerns and uncertainties to social and economic norms and structures.

Right now, AI is literally challenging the essence of innovation. A standard mechanism of invention starts with research and ends up with patents and benefits. But AI seems to do everything by itself, so it can lessen the incentives for innovations and question the beneficiaries.  Clearly, AI identified existential questions about the core tenets of the patent system including ownership, inventorship, and infringement. And the DABUS case aggravated such an issue.

What is to be done? 

The degree of depth of this discussion varies in different countries. But the question of what to do next is acute almost everywhere and there is no clear answer. The graph below shows different AI policies that leading economies utilize. The policies are limited to database rights and national AI strategies. One of a few good practices is a series of consultations happening in the UK on how to regulate AI-devised inventions.

Policies related to AI

Source: based on WIPO

Today countries need to unambiguously state at least what AI can and cannot do. There are four possible scenarios for lawmakers depending on national priorities:

  • No legal changes. This “Let-it-resolve-by-itself” approach is definitively not optimal, and we are not calling for this, but yet it can help to avoid major shakeouts in the patent system.
  • Prohibit AI to be an inventor. With this approach, one does not interfere with human R&D, but it can decelerate AI advancements.
  • Allow AI works and do not protect them. This scenario may facilitate innovation in AI technology and promote its use for the public good. But who will bear responsibility for possible negative consequences of AI inventions then?
  • Allow and protect AI works. Protecting AI works may boost the sector and promote new AI inventions. Though it may be at the expense of human R&D and devalue human creativity. Should it be protected for a limited period, say, one year? Or should it be allowed only for socially significant sectors, such as education and health?

We look forward to your views on these scenarios and the existential question of whether AI can indeed sway unquestioning only human creativity tenet. Clearly, there is a sense of urgency for developing countries to choose the appropriate scenarios for their national agenda. AI inventions may have a critical impact on national innovation potential and economic growth. If no action, AI may amplify digital divide globally.


Igor Slabykh

Lawyer, Director, First Priority Consulting LLC

Yaroslav Eferin

Digital Development Consultant

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000