Inspired or copied, now with robots. Protocols for the Use of Artificial Intelligence and Intellectual Property

The use of Artificial Intelligence -AI seems to be reaching several aspects of our lives and has us all thinking about the consequences and challenges that its coming of age implies.

The concept of IA is not new, but now that it is closer to us, some of us have already begun to look more closely at the ways in which AI can affect, for better or worse, the way we work and live.

The first thing we must say is that AI does not “create”, at least not yet. The basic concept for AI is “machine learning”, that is, feeding information to a machine so that, according to what we have fed it, it knows how to react to a situation, develop content, or respond to a question.

And that is the issue I want to talk about in this opportunity. The materials, content, information, and samples with which the machine is trained may include works protected by intellectual property. How do we handle this situation? Should this be allowed?

There have already been cases in which visual artists have complained that AI platform developers have used artwork created by these artists to train their machines. Getty Images, a major photography agency, claimed the use of its photographs (copyrighted works) used for training Stable Diffusion, a generative AI model that produces unique photorealistic images from text messages and images.

And then we come to the fine and delicate border of copy versus inspiration. In copyright, there is the concept of a derivative work, that is, a new work based on an existing one. The rights of the initial work are reached to the new one. But there is also the concept of “fair use”, that is, the exceptions under which a creator can use the works of third parties to create her own (under certain conditions, of course).

In the case of Stable that we talked about above, the company offered that, in the next version of the tool, creators can request that their works be removed from the catalog with which the machine trains. This is not sufficient according to current intellectual property rights protection standards. Saying “I'm going to use your work and if you don't agree I'll stop doing it” seems insufficient to me.

In this context, we then have the potential drawback that the content generated by AI for the next marketing campaign, the technical content harvested in an Internet search for previous works or state-of-the-art material used for new development, the texts of a conference, all of this could have an owner.

Of course, this problem is not new and has not been created by AI. The unauthorized use of materials (inadvertent or intentional) has always existed. But what can be misleading is to consider that since the materials or contents and results of an investigation were “created” by a machine, they are new and innovative and therefore can be used without looking a little further.

This doesn't mean not exploring how AI can help your business, and no matter the size of your endeavor. In fact, AI can be very efficient in small companies and ventures that do not have so many resources or in large multinationals in which the volume of operations and processes involves titanic efforts.

At CMS we believe that these tools are here to stay and will change many aspects of the way we work. We have already supported companies in the development of Artificial Intelligence management protocols that contain provisions related to intellectual property, ethical management, and personal data, among others.