The world is currently going through a technological revolution, and artificial intelligence (AI) is one of the major players. AI has been revolutionizing everything from healthcare and finance to transportation and entertainment. As an AI model developer, building AI models usually involves more than just writing code; it involves research, development, and testing. As a result, protecting that intellectual property is of utmost importance. However, navigating the legal landscape surrounding AI model intellectual property can be confusing.
In the world of AI model intellectual property, rights arise from two main sources: patents and copyrights. A patent is a legal document that gives the holder exclusive rights to an invention. In contrast, a copyright is a legal mechanism that protects creative works such as music, literature, and software.
When it comes to AI models, obtaining a patent is difficult, if not impossible. This is primarily because patent offices typically require that the invention it covers be non-obvious, novel, and useful. However, AI models almost always utilize pre-existing algorithms. For this reason and others, there are currently very few AI model patents.
On the other hand, when it comes to copyrights, the landscape is more navigable. Although it is an ongoing debate about whether or not AI models are creative works, software that runs on a computer is ordinarily considered a creative work, and as such, is generally deemed eligible for copyright protection.
But it is worth understanding that not every computer program is eligible for protection under copyright law. In general, programs that are created “independently” and deemed “original” are subject to copyright protection. This means that copyrighted works need to be more than just a sequence of characters and symbols, but they must have creative human input or involve original authorship. Therefore, it’s important to take the necessary steps to ensure protection of AI models.
One way to do this is to ensure that AI model ownership is properly attributed and recognized through the use of licensing agreements. A licensing agreement outlines how an AI model can be used. It typically includes information on the scope of rights that are granted, the fee structure, any conditions or restrictions on use, and any exclusions on the rights that are granted.
Licensing agreements exist in different forms, some of which are more restrictive than others. For instance, a proprietary license grants the licensee very limited rights in an AI model. Alternatively, an open-source license grants the licensee extensive rights in an AI model.
One more distinction to keep in mind is the difference between exclusive and non-exclusive licenses. Exclusive licenses grant the licensee the right to use an AI model to the exclusion of all others, including the original owner. On the other hand, non-exclusive licenses grant the licensee the right to use an AI model, but not exclusively.
When it comes to AI models, several companies, including IBM, have taken a unique approach to licensing. IBM specializes in providing wide-ranging licensing opportunities to AI models. For instance, IBM offers standard licensing, i.e., a non-exclusive license on a per use basis. An example would be using an AI model for chargeback detection and prevention within a specific point of sale system. IBM also provides broader “partner” licensing agreements that give businesses unrestricted use of the AI models and can take shape as a service to clients or as end-products for sale. As a result, IBM leverages its AI licensing model as revenue generation, due in part to their massive M&A activity in AI and related technologies.
In conclusion, AI model intellectual property is a complex subject, and licensing is one way to protect those valuable IP assets. However, AI model ownership and licensing agreements remain a nascent area of the law and can vary from jurisdiction to jurisdiction. Therefore, it is important to work with competent legal counsel when dealing with these matters. As AI technologies continue to advance and become more commonplace, the legal landscape and risk considerations will dramatically change.