Monday, February 27, 2023


The headlines these days are full of stories about artificial intelligence (“AI”). It seems AI has gone from science fiction to reality in a very short period of time. From Open AI’s ChatGPT app to Google’s MusicLM music creation tool, the future is now. A few decades back the struggle over digital rights and copyrights was a sideshow, of concern only to a relatively small slice of the population. Now, however, everyone is online, which means that even if you don’t consider yourself a creator as such, material you write, post or share could become part of an AI engine and used in ways you never imagined. To wit, several of my clients have recently raised concerns about the threat of AI, and the threat AI may pose to authors, artists (both fine and graphic artists) as well as songwriters and musicians… so let me try to break it down.

The tech behind AI is a complicated, but at its most basic AI engines ingest massive data sets, which they then use to train software that can make recommendations or even generate code, art, or text. Sometimes the engines are scouring the web for these data sets so they can learn what is on a webpage and catalog it for search queries; other times, AI engines have access to huge proprietary data sets built in part by the text, photos, and videos users have posted on their respective platforms. Unlike the music piracy lawsuits of decades ago, the AI engines are not making copies of the data they use and distributing them under the same name. The legal issues, for now, tend to be about how the data got into the engines in the first place and who has the right to use that data.

AI proponents argue that AI engines can learn from existing data sets without permission because there is no law against learning, and that turning one set of data into something entirely different is protected by the law, as affirmed by a lengthy court fight that Google won against authors and publishers who sued the company over its book index, which cataloged and excerpted a huge swath of books. Those against AI argue that if AI is going to use original material created by others then the AI engines should get a license from the original author or creator, or from whoever owns the copyright. Otherwise, the AI engines are violating the rights of the authors and creators of the underlying materials and, even if the result is not an exact copy of the underlying work (think “sampling” in music), the resulting AI product would not exist “but for” the underlying work and the use of the AI work competes with the artists or creators’ ability to make a living.

As of now, the US Copyright Office has stated that it will not accept a copyright registration on any work of authorship created using AI. However, there is still the issue of whether the resulting AI creation infringes the copyrights of other authors when it incorporates the work of the original authors into a new AI work or some sort (regardless of how remote or miniscule the resulting use in the new AI-generated work). Using the MusicLM system as an example, the Google researchers noted that MusicLM did have a tendency to incorporate copyrighted material from training data into the generated songs. During an experiment, they found that about 1% of the music the MusicLM system generated was directly replicated from the songs on which it trained (a threshold apparently high enough to discourage them from releasing MusicLM to the public in its current state).

Assuming MusicLM or an AI system like it is one day made available, it seems inevitable that it will give rise to major legal issues, even if the systems are marketed as tools to assist artists rather than replace them. AI music generators violate music copyright by creating a mosaic of coherent audio from prior works they ingest in training, thereby infringing the US Copyright Act’s reproduction right. Similar concerns have been raised around the training data used in image, code and text-generating AI systems, which is often scoured from the web without creators’ knowledge. New works generated by an AI system would arguably be considered derivative works, in which case only the original elements of the AI-generated work would be protected by copyright (although it is unclear what might be considered “original” in such a context).

It might not be long before there is more clarity on the matter. Several lawsuits are making their way through the courts which will likely have a bearing on music-generating AI, including one pertaining to the rights of artists whose work is used to train artificial intelligence systems without their knowledge or consent.

Current AI platforms are crude and imperfect, but by design AI improves and gets “smarter” the more it is used. AI may still be in the early horse & buggy stage - but AI could quickly advance to the Tesla level. Authors, creators and artists beware!


Wallace Collins is an entertainment lawyer and intellectual property attorney based in New York. He was a songwriter and recording artist for Epic Records before receiving his law degree from Fordham Law School. 



No comments:

Post a Comment

Note: Only a member of this blog may post a comment.