Will our inventions soon be able to come up with inventions of their own, and can one decision deprive European entrepreneurs of the right to prohibit when it comes to their inventions? European Patent Attorney Marjut Honkasalo explains why artificial intelligence and the patenting of simulation methods will be whipping up a storm in the IPR teacup in the coming year.
From invention to inventor? The discussion around artificial intelligence heats up
The more patent and trademark applications are filed in a new area, the more necessary it becomes to shake up traditional application practices. In early 2020, the notion of whether artificial intelligence can be an inventor caused a stir in the IPR world. Discussion was raised by the patent applications filed by Stephen Thaler, the leader of a British research team. The applications were rejected because the artificial intelligence DABUS had been named as the inventor.
“Thaler appealed the decisions and forced the offices to rethink whether artificial intelligence can be an inventor. Britain ruled against the idea, but the United States and the European Patent Office EPO have yet to present their final positions. The ruling affects both the applicants and holders of IP rights”, says our European Patent Attorney Marjut Honkasalo.
The offices in different countries, including the EPO, will determine whether we need changes to the patenting practices, or even legislation, concerning artificial intelligence. Of the two main points of contention, the first relates to the criterion of creativity.
“How do we determine when something is created by artificial intelligence on its own and when only with the help of artificial intelligence? For now, it is a completely sci-fi concept that artificial intelligence would create something truly new independently. Artificial intelligence efficiently processes the huge amounts of data fed to it, but it still lacks creativity”, Honkasalo estimates.
The second issue relates to the transfer of rights. Unless legislation is amended, artificial intelligence is not a legally competent entity that could transfer the rights to an invention.
“In Finland, an employer acquires the rights to a human inventor’s invention under the Act on the Right in Employee Inventions, and a party outside an employment relationship or even a CEO can transfer his or her rights to a company. People and companies are legally competent entities that can transfer rights to each other, but artificial intelligence is not.”
It would be easiest for applicants of intellectual property rights if applications related to artificial intelligence were interpreted in the same way around the world. The authorities are working towards compatible interpretations, but in practice, it is not easy to reach an agreement.
“Gazing too much at one’s own navel can ruin a unified line if each country thinks that their position is the only correct one that others should follow as well. When planning protection, it is worth remembering that artificial intelligence is at present thought to be smarter than it actually is. AI solutions can be patented, but as long as artificial intelligence is unable to create something truly new on its own, it can hardly be an inventor”, Honkasalo says.
Majority of simulation methods may be left without a patent in the future
3D printing is a newcomer of the digital age, but simulation is gaining popularity alongside it. Computer simulation is increasingly being used in, for example, various digital twins, engine modelling, drug molecule development, and structural design. The most critical problem lies not in simulation itself, but in the criteria for patenting it.
“The EPO’s extended Board of Appeal is currently considering what simulation methods can be patented in Europe and which features of simulation can be taken into account in the assessment of patentability. It depends on the position of the EPO whether a large portion of simulation methods will be left without a patent. In the worst case, it will become more difficult to obtain a patent in Europe, and real inventions will be left without patent protection”, Honkasalo says.
The debate was sparked when the EPO’s Board of Appeal considered the simulation of crowd movement in a planned building and wanted to deviate from the ruling issued by another Board of Appeal regarding the electronic circuit noise simulation of semiconductor manufacturer Infineon Technologies. According to its ruling, a simulation method does not have to involve the manufacture of an end product if it is limited to a specific technical use or contributes to a specific technical implementation. According to the second Board of Appeal, computer simulation alone is not yet sufficient to establish the technical nature of an invention, but rather a simulated end product must be included in the scope of protection.
“In modelling and structural design, this would mean that patent protection could only be obtained if the simulation is actually linked to the construction of a building. In simulating an explosion, you would have to include the detonation of something in the scope of protection in order to be granted a patent. That would be silly.”
“The ruling would also be problematic in terms of patent infringements because, for example, in construction, a patent infringement would require both simulation and construction – planning alone would not be enough to infringe a patent. In the case of medications, an infringement lawsuit could only be brought once the medication is in production, which takes time. Patents could become practically worthless”, Honkasalo says.
EPO’s decision may take away the right to prohibit in one’s own market area
If the EPO’s decision were to conform to the latter position, it would have a negative effect on both pending patent applications and those sought in the future. According to Honkasalo, the decision would be particularly unfortunate for local operators in Europe.
“The possibility of patenting a simulation method in China and the United States is small comfort for a European operator that would be unable protect the simulation alone and prohibit competitors in its own main operating area. If this position were the final ruling, the majority of the applications currently pending in Europe regarding simulation are unlikely to be accepted. This means that the patent applicants would lose both the money and time invested.”
The decision of the EPO’s extended Board of Appeal may be issued at any time as no specific deadline has been set for it, and questions were submitted to the Board already in early 2019. The Board of Appeal may not necessarily answer all the questions put to it either.
“The EPO wants to be an authority to which applications are submitted. If the patenting criteria are too strict, the EPO excludes a huge area of patenting from its scope, which would be a bad option for them too. That is why the probability of a, in my opinion, silly decision is quite low”, Honkasalo says.