This article is the first in a series exploring legal issues impacting artificial intelligence (AI) use and development.
Even though the capabilities of AI are still in their infancy, the use of AI is already pervasive in all industries. So, it should not be surprising that AI is testing legal norms, and we should expect to see many legal actions focused on AI usage. One of the initial cases is New York Times v. OpenAI and Microsoft filed last December.
As background, AI constructs algorithms designed to mimic human thought processes and decision making. A human mind is constructed with a database (knowledge) and a processor (the brain’s functions of analysis, judgement and creativity). In order for AI algorithms to do their magic they need data. A brain without information will not function. Once the data is fed into AI, then AI goes through a continuous development process of creating algorithms from the data and learns how to use such data.
OpenAI has created an AI system and Microsoft, as an investor and partner, has been incorporating the OpenAI functionality into its products.
One major source of gathering data has been publicly available news, like the New York Times. By digesting NYT’s content, OpenAI is able to create NYT’s-like articles, that may even sound like the writing style used by many NYTs journalists.
The NYT has claimed that the digestion process and subsequent use of the NYT’s content constitutes copyright infringement. The defendant’s claim that their use is “fair use” which is an exception under the U.S. Copyright Act to what would otherwise constitute copyright infringement.
These are all issues of first impression. The courts will eventually sort all this out. In the meantime, at these early stages, AI developers and users need to assess the risks of law suits.
However, what in all likelihood underlies the NYT’s suit, really in plain view, is not copyright infringement and the alleged threat to journalism that the NYTs claims in its complaint, but rather the immense profit that OpenAI/Microsoft expect to make from AI.
In short, the NYT is probably fine with being used by AI as a content source for teaching the AI functionality, but in this age in decreasing revenue earned by traditional news sources, the NYT, as a premier news source, wants major compensation for use by AI of its content. The complaint even mentions that the parties were attempting to negotiate a settlement, but negotiations broke down. The settlement probably did not seek damages, but instead royalties. How do I know that? Not through any inside information, but because last July, OpenAI announced a licensing arrangement with The Associated Press, so you see this is part of the business model.
In short, the lawsuit by the NYT is likely not an effort to assert their legal rights as much as an effort to pressure OpenIA into making a lucrative licensing arrangement.