Patent law is predicated on the premise that inventors are humans; it currently struggles to accommodate a machine inventor. As patent applications designating an AI system as the inventor have been filed in over 100 countries, courts throughout the world are currently grappling with this issue.
In the year 2020, a machine-learning system assisted scientists in developing a strong antibiotic that is effective against several infections.
Artificial intelligence is also utilized to help in the creation of vaccines, medicinal design, material discoveries, space technologies, and ship design. Within the next several years, various inventions may include AI. This is one of the greatest dangers to patent systems ever encountered.
Multiple organizations are holding public consultations on artificial intelligence (AI) and intellectual property (IP) legislation, notably in the United States, the United Kingdom, and Europe.
If courts and governments rule that Artificial intelligence AI-generated ideas cannot be copyrighted, the repercussions might be enormous.
Funders and corporations would be less motivated to support meaningful research with AI innovators if the potential return on investment was low. The creation of beneficial and life-saving inventions may not occur.
Instead of attempting to force antiquated patent rules to adapt to new technologies, we recommend that national governments establish AI-IP laws to protect AI-generated innovations.
In addition, nations should form an international convention to guarantee that these laws adhere to standardized principles and that any conflicts may be addressed expeditiously. Researchers are required for both phases.
‘Who’, not ‘What’
The drafters of the world’s first patent law, the Venetian Patent Statute of 1474, did not anticipate the possibility of self-creating machines.
The 1883 Paris Convention for the Protection of Industrial Property, which formed the underpinnings of the worldwide patent system, did not account for them either.
In 1994, when the World Trade Organization finished its Agreement on Trade-Related Aspects of Intellectual Property Rights, AI-generated ideas were nearly unheard of (TRIPS). International patent standards are mandated by the 1883 and 1994 conventions.
The TRIPS agreement protects “all inventions, whether goods or processes, in all technological domains, so long as they are fresh, innovative, and capable of industrial application.”
In its language, the phrases ‘inventions,’ ‘new,’ ‘inventive step,’ and ‘capable of industrial application’ are specialised terms, each of which has a legal definition. Essentially, an object is not patentable if any one of these conditions is not satisfied (see “What is patented?”).
New technologies have previously posed a threat to the system. Numerous high-profile cases have examined the patentability of genetic sequences, human-made biological beings, and other items.
The primary legal issue in these instances was whether or not they constitute innovations. In 2013, the US Supreme Court ruled that isolated human gene sequences are not patentable since genetic information is a product of nature and not a human invention.
AI-generated inventions provide a novel challenge to the patent system since the question is ‘who’ rather than ‘what’ was produced.
Patent registration offices were first and most urgently concerned with whether the inventor needed to be human. If not, it is feared that AIs would eventually be so abundant that patent applications for their discoveries will overwhelm the system.
An another difficulty is even more basic.
When an invention is judged “non-obvious” to a “person skilled in the art,” this constitutes a “inventive step.” This hypothetical individual possesses the average degree of expertise and general knowledge of a typical specialist in the applicable technical subject.
If a patent examiner determines that the invention would not have been obvious to this hypothetical individual, the invention moves one step closer to being granted a patent.
However, it is unclear how a human patent examiner could determine if an AI’s innovation was apparent if AIs became more informed and skillful than all humans in a subject.
An artificial intelligence system designed to analyse every material released on a field of technology before inventing would have a significantly wider knowledge base than a human. Considering all information, nearly everything would appear clear.
If everyone in the future had access to these AI tools, the ‘inventive step’ requirement of patentability would be almost hard to meet, and essentially nothing would be patentable. A comprehensive rethinking would be necessary.
AI-IP Law
Modifying current legal safeguards has the danger of creating ambiguities, hence full law change is preferred. The optimal answer would be for governments to create a unique kind of intellectual property known as a sui generis law.
These custom-made rules are intended to protect sorts of creative production that are not covered by the “big four” intellectual property (IP) doctrines: copyright, industrial designs, trademarks, and patents.
In certain jurisdictions, they currently stimulate and safeguard investments in circuit design, novel plant types, and databases.
Some critics may object to the proliferation of topic-specific IP forms. However, a separate AI-IP concept has the benefit of being adaptable to the specific conditions in which AI creativity emerges.
Legislators might determine, for instance, that if AI-IP is easier and quicker to create, it should be protected for a shorter period than the typical 20-year duration of regular patents.
This would encourage others to build upon inventions after the expiration of the patent. And although patents are normally granted to the inventor, legislators may opt to share the profits from an AI-generated innovation differently, say among the AI developer, the person leading the AI, and the owner of the data used to train the AI.
International Treaty
Countries that implement AI-IP would be likely to attract research and development investment. On the other hand, royalties connected to the use of innovation may reduce its accessibility.
A similar problem exists today with pharmaceuticals and vaccines: patents can attract the investment necessary for their development, but individuals in countries who cannot afford the items or pay royalties to produce them suffer.
Approximately 16 percent of persons in low-income nations have gotten at least one dose of COVID-19 vaccines approximately 18 months after the vaccines initially became available.
In the meanwhile, some nations with better incomes are giving fourth dosages.
In constructing AI-IP, a balance must be struck to prevent the reproduction of such inequality.
