Are AI and intellectual property ¿“frenemies”?

By: Natalia Vera Matiz

Natalia Vera Matiz, partner at Vera Abogados Asociados, explores the two main concerns surrounding the coexistence of AI and IP.

When we discuss AI and its interaction with intellectual property, it seems that the first constitutes a threat to the second. To assess this preconception, it is important to consider different points of view.

The issue at hand is not minor, especially considering that many authors nowadays use AI in their creative process. This raises two concerns: first, determining the extent of human intervention required to consider a work created with the assistance of AI as eligible for copyright protection.

The most straightforward conclusion for this issue is that works created solely by AI are not entitled to copyright protection. However, this approach could hinder the advancements of the technology industry and investment in this field of the economy. Just as the cinema industry has benefited from changes in copyright legislation that allowed for increased flexibility in the use of copyrighted materials, encouraging innovation in filmmaking, the attitudes towards AI authorship could turn the other way.

The second issue concerns the use of materials for training AI. Some advocate for the free use of materials to train AI, arguing that free speech and scientific research are at stake because AI can constitute an important advancement in both fields. The counterargument concerns the lack of remuneration for and authorization from the owners of materials used in training AI models.

Lawmakers and judges need to consider implementing policies that provide remunerations alongside setting exceptions for obtaining a license or authorization from the owner to use protected materials in the process of training AI models. These subjects have been taken into account in recent cases, including The Times v. ChatGPT. We should also look to other countries; Japan, for instance, has proposed a very flexible draft policy to allow for the broadest possible training of AI with copyrighted material.

In order to foster the development of generative AI and the technology in general, it’s time to ask the right questions and avoid arriving at premature conclusions, such as ruling out the protection of works produced by AI altogether! We will not be able to train AI to its fullest potential without copyrighted material; it’s time to think globally once again, achieve a general consensus before it overwhelms us all, and think with the mindset of the original purposes of IP.