Artificial intelligence is not a new concept, but only recently has AI with regard to patent applications been given more prominence. Now the European Patent Office has rejected two patent applications which designate the inventor as a non-human entity. According to the patent community, future AI developments could also impact the established principles of patent law.
6 January 2020 by Amy Sandys
The EPO has refused two patent applications following non-public oral proceedings in November. The applications were rejected on the grounds that the applications list AI, rather than a human, as the inventor. There is concern that innovative business models which seek to employ AI will be restricted by laws of inventorship. However, the global patent system might also have to adopt a more flexible approach in how it assesses the relationship between AI and patents.
The decision, released at the end of 2019, states that two patent applications are invalid. This is due to the inventor not being listed as human being, a principle set out by Article 81 and Rule 19 of the European Patent Convention.
The two patents, EP 18 275 163 and EP 182 751 74, concern a fractal beverage container and fractal light signals respectively. The fractal light signal embodies a device to attract attention during search and rescue operations. Stephen Thaler, a doctor of physics, created the AI. Its name, ‘DABUS,’ stands for ‘device for the autonomous bootstrapping of unified sentience.’
Neural networks, computing systems based on biological neural networks found in sentient beings, make up the technology. The inventor is cited as DABUS on both applications, which were also submitted in the US and the UK. However, the respective patent offices also rejected these. Patent offices in Germany, Taiwan and Israel have also received the applications, and more will follow.
Thaler argued that the inventions had acquired the right to a European patent. This was on the basis that the AI inventor was his successor of the inventor title. However, the EPO rejected this argument. The decision has not come as a surprise to patent attorneys.
Pawel Piotrowicz, UK and European patent attorney at Venner Shipley, says “This decision by the EPO confirms that an inventor must be human, a view shared by the UK and US patent offices. However, AI technology is continually developing and so the issue of AI inventorship may need to be reviewed again in the future.”
Piotrowicz notes that an AI system being viewed as an inventor could affect the assessment of inventions. This is particularly in relation to obviousness.
The decision has let to much commentary on social media and online forums. Some in the patent community are concerned that the EPO decision could have consequences for businesses developing inventions based on AI principles. Other companies are increasingly turning to AI to speed up the invention process, as it cuts out much of the human labour cost.
Analysts say that AI is positive for rapidly-changing fields such as healthcare. However, it leads to difficult questions concerning who owns the technology.
The EPO announced that it would releas a reasoned decision in January 2020. The decision must be published before an appeal is possible. However, the UK decision is currently already under appeal.
Update: the EPO has now released the reasoning behind its decision. Firstly, the EPO considered that interpretation of the European patent system framework leads to to the conclusion that the inventor of a patent must be a ‘natural person.’ The office also considers that the notion of a ‘natural person’ is internationally applicable; it argues various courts have handed down judgments to this effect.
Finally, the EPO argues that designating a machine inventor with a name ‘does not satisfy the requirements of the European Patent Convention.’ The full reasoning can be found here.
For DABUS
Williams Powell (London): Robert Jehan (lead)
This article was updated on 28/01/2020 to reflect the latest developments at the European Patent Office.