The growing role and potential implications of AI on patents
As artificial intelligence systems start to co-author patents, how they are treated as part of the organization is going to be a challenge for companies using them for patents.
AI is without doubt the latest buzzword in town. Although it’s not as recent a phenomenon as one thinks, it has definitely penetrated our lives in the recent years more than ever. Be it Google assistant, Amazon’s Alexa, or Apple’s Siri – all of them are powered by a backend AI engine, which is ever learning and ever improving.
Although, AI is penetrating every field now, pharma was one of the earliest adopters of the technology utilizing the AI engines in drug discovery, diagnosis of disease. In 2017, the researchers at Stanford university trained an AI engine using 129,450 clinical images, and then tested its performance against 21 certified dermatologists in identifying Skin diseases. The AI was proven to be as efficient as human dermatologists in recognizing skin cancer. When it comes to patents and AI, there are several issues as discussed in following paragraphs.
Patentability of AI based inventions
As much as it is complex in nature, and ever learning, and ever improving, AI is a software, and software patentability has been frowned upon by law across the globe. In Europe, Article 52 of the European Patent Convention clearly states that “The following in particular shall not be regarded as inventions: schemes, rules and methods for performing mental acts, playing games or doing business, and programs for computers”.
Similarly, the Indian Patent act in Section 3(k) also states the computer programs as not being patentable per se. Although, the law seems to be united against the patenting the AI, but there have been workarounds for patenting the software, for example, in India, the software patents have been allowed in case they are tied up with a hardware, for example, the computer processor on which the software executes, so AI related inventions will not likely be very different.
AI as an Inventor
Perhaps, the most interesting issue in the patenting industry is of the AI being a named inventor of the algorithms it develops. DABUS (“Device for the Autonomous Bootstrapping of Unified Sentience”) is an AI system developed by Dr. Stephen Thaler, which was also named as an inventor in two patent applications on the technology that was developed by the AI. The patent applications, with DABUS as inventor, were filed in US, UK and European Patent Office.
All the three patent offices rejected the patent applications that named DABUS as inventor. The key grounds for rejecting the patent applications were that the laws, as written, envision a natural person to be an inventor and not an AI.
Clearly, there’s a gap in the current legislations which the law makers across the globe need to ponder over and address in case AI systems were to become inventors.
But things do not stop there, in case an AI invented something, who owns the technology and the intellectual property generated thereby?Most of the organizations own the intellectual property generated by the employees, as they are working for the organization for emoluments. Is it safe to consider that DABUS, was working under the supervision of Dr. Stephen and his organization, therefore, the intellectual property belongs to the organization? That would that make DABUS an employee of the organization, giving the AI legal rights.AI being an ever learning system, what if DABUS further evolves and says no?
AI assisted patent intelligence
AI has impacted many industries, and Patents is no different. The major benefit of the AI would be to make the process more efficient and faster across all levels – the patent searching, patent examination and grant process, and even patent licensing. The patents data is very structured across the globe – there is a defined way to draft a patent, there are defined sections in patents that contain specific information.
If AI systems can help consume unstructured data and observe trends from it, absorbing patents data to identify patterns will be a relatively easy task for an AI engine. AI is already integrated in may patent searching tools today, and are making the searching process more and more efficient.
For the same search query, a search engine backed by AI results in less false positive results.AI based translation engines are making more and more patents accessible to public by translating patents to English. The Indian Patent Office is also interested in leveraging the AI based engines to make the patent granting process more efficient.
From the standpoint of patents, the AI is still in the nascent stage and we are yet to witness the full extent of the benefits it can provide. There is no denying that the AI systems will prove to be tremendously beneficial to everyone by making the process more efficient and faster. Also, the legislation across the globe seems to be the lagging behind, as most of the patent laws today were written in an era when AI was just a science fiction, which will need to be updated to match the recent phenomenon.
The author is Vice President, Effectual Services- Advisory firm that offers IP support solutions