Per Law 360, a service which updates attorneys on major trends in the law, the often-discussed issue of whether AI is stealing the intellectual property of others, as that property is appearing in publicly available sources, is the subject of growing and important litigation, with no clarity as to which direction the law will take.

There have been over thirty cases, all in one of Massachusetts, California, Delaware and New York.  They arise across the spectrum: some allege misuse of text, some the theft of images, some the theft of music.  Three cases have been decided; one held that a company infringed on the copyrights of Westlaw (a platform of information for attorneys) while two others found that it was fair use to use public information in training AI.  The discussion of fair use in these cases did not result in a common interpretation and, according to speculation reported in Law 360, it is expected that it will take three to five years to arrive at an understandable body of legal interpretation.  By then, some standard basis for contractual settlement of the issue as between parties may also develop.

Finally, there is an esoteric discussion as to whether this litigation really is measuring “the heart of what makes us human” as it asks a court “whether AI models can legally be used to create art that replaces human works.”  I venture to suggest that this artistically sensitive view of the issue is far less important to people who pay money to institute lawsuits than the possible economic return upon any successful legal claim.  But on one level, consider this: what is the difference between a person reading available literature and writing about the subject based on that reading, vs a machine doing exactly the same thing?