Intellectual property litigation in relation to Artificial Intelligence (AI) and machine learning
The litigation issues involving patents and trade secrets related to artificial intelligence (AI) and machine learning is going to increase as they are being integrated into an increasing number of products and services. Some of the legal experts expressed their views in recently published interviews where they reflected on the evolution of AI and machine learning in a broad range of industries, their likely impacts on intellectual property (IP) protection, the preparedness companies should do now to prepare for future legal issues.
Why patents haven’t come into litigation yet?
According to one of the industry expert, AI and machine learning are very new and that’s why patents haven’t come into litigation yet. He says an application takes about three years to get through the U.S. patent office to issuance and thus one may start seeing enforcement activity ramp up in another year or two.
Another expert says that some clients have the realization that several legal issues may be coming down the road. She opines that new issues in patent law may include who is the inventor if an artificially intelligent system discovers a new innovation. According to her, one of the major concerns is how liability will be ascertained if an artificially intelligent system makes a decision which causes monetary damage or harm. Another concern is spelling out in agreements, responsibility for the decision-making or results they get from automated or AI systems clients may be using now or in the future. The ownership of data, information, or results that may be generated by the artificially intelligent system is also a concern for the clients which will be required to be spelled out in an agreement.
Ownership of the outputs of AI systems that are able to engage in the act of creation is going to change to the future litigation scenario. The act of creation can involve anything with potentially traditionally copyrightable subject matter, including music, or solving particular problems or creating engineering solutions that might be patentable subject matter.
It is opined that the legal risk involving AI and IP is not just for the tech, automotive, and transportation companies but will extend to industrials companies, energy companies, agricultural, and chemical companies, as they will also enter into development and use of artificially intelligent systems.
On the difference between machine learning and AI, it is said that machine learning is a subset of artificial intelligence which involves a large set of training data. On the infringement issues companies may face if they use AI or machine learning in their products, it is opined that one of the main issues is determining whether to seek patent protection or keep a machine learning innovation as a trade secret. At this time there is no easy way to determine if one’s machine learning patent is being infringed. There is almost no way to determine other machine learning process and algorithm without suing them, which can be very expensive and risky. So the client has to decide whether to file for a patent or go through all that expense of trying to figure out if there is infringement.
Patents and Trade secrets
On the distinction between patents and trade secrets, one of the panelists said that trade secret is something that is defined as information that is not public and that has independent economic value from the secret not being known and it lasts as long as it’s kept secret. The infringement of trade secret depends on the wrongful taking of it. But the Patents are protected for a limited period of time and its disclosure should enable somebody to make and use the invention. The inventor, in exchange for disclosing the invention, gets the exclusive right to practice the invention for a limited period of time during which its more of a strict liability right, where if someone wanders into the space protected by a patent becomes liable for infringement.
On the types of information clients should include in their AI-related patents, it is observed that a simple thing which is being done for a long period of time can’t be patented unless the process of machine learning is defined. If the client has to disclose the specifics and detailed implementation choices that client has selected, and it can start to limit the value because there may be many different ways a particular problem can be tackled. The panelist is of the view that the question from the perspective of inventors is whether the inventor wants to disclose how the AI goes about doing things, although it all happens inside the algorithm, but the flip side of this is that description in the patent is necessary in determining infringement.