Are You Ready for Artificial Intelligence?

Impact on (IP) Law

The use of AI by businesses and individuals has already raised several IP questions. For example, courts around the world have been called on to assess whether an AI can be the name inventor on a patent or the author of a copyrighted work. So far, they have generally answered that question negatively.

There are also several cases pending concerning the use by AI developers of IP-protected works in training their algorithms, whether such use is fair use, and whether the resulting product is considered a derivative work.

Ms. Loadman says this “infringement risk” is something that everyone who uses generative AI needs to consider, because the scope of protection and liability are far from clear. “How are the AI tools trained? Where is that data coming from and who owns the content? We don’t know if it’s fair use,” she says.

Writers such as John Grisham, George R. R. Martin, and Sarah Silverman, as well as Getty Images and the New York Times, have brought cases against generative AI companies, and the outcome of these cases—which have been filed in both the U.S. and Europe—should help to clarify the law. “I’m hoping we get good guidance that makes sense for the direction the technology is heading in,” says Ms. Loadman.

Meanwhile, there are legislative initiatives in progress. The EU AI Act, which was recently adopted by the European Parliament, includes provisions requiring generative AI developers to be transparent about data sources and to comply with copyright law. In the US, Representative Adam Schiff introduced the Generative AI Copyright Disclosure Act in Congress on April 9. This would require AI companies to submit copyrighted works used in training to the Register of Copyrights before releasing new systems.

AI’s impact on trademark practice is likely to be significant too. For example, AI tools can be used to generate brand names, logos, product designs, and advertising campaigns: one example of the latter is an ad campaign for animal foods that uses AI to generate flawless dog pictures better and more easily than using real animals.

AI has enormous potential to make processes such as trademark searching and, in particular, image searching, more efficient and accurate, and is already being used to identify counterfeit goods.

However, Prof. Lim points out: “In trademark law, AI-generated marks could increase the risk of confusion and dilution.” He has written a paper, Computational Trademark Infringement and Adjudication, that addresses how AI could both challenge and assist in trademark analysis and enforcement.

As with other technological developments, such as online commerce and social media, a danger for brand owners is that wrongdoers may be a step ahead of them. For example, highly accurate and realistic deepfakes could be used to indicate an association or endorsement that does not exist, leading to potential damage to a brand or commercial partner such as a sponsor or influencer. Such activity could raise questions involving laws on publicity and privacy as well as IP. “Criminals are already playing with this technology,” warns Ms. Razavi.

“Given the global nature of both AI and IP, international cooperation and harmonization of laws will be crucial in effectively addressing the challenges posed by AI to copyright and trademark laws. For example, AI could bring about more consistency in trademark registries and enforcement practices worldwide as global databases become more integrated and accessible for AI analysis,” says Prof. Lim.

“Law will adapt to technologies and the issues will become clearer,” says Ms. Loadman.

Prof. Lim agrees, adding: “The rules on inventorship and authorship are robust. We will likely have clearer guidelines around infringement and fair use this year. My concern is whether the international legal order will be able to navigate these changes given the tugs of nationalism and polarization on a scale unseen since the years preceding the two World Wars.”