Feb 22, 2024 12:30 AM
https://arstechnica.com/tech-policy/2024...st-openai/
EXCERPTS: . . . So far we’ve focused on the legal vulnerabilities of AI companies. But the defendants could have one big factor working in their favor: judges reluctant to shut down an innovative and useful service with tens of millions of users.
Judicial opinions are framed as technical legal analysis, but they are often driven by a judge’s gut feeling about whether a defendant has created a useful service or is merely seeking to profit from the creative efforts of others.
[...] A judge who feels a product is valuable will look for ways to allow it to stay in business. On the other hand, a judge who feels a product unfairly profits from the creative work of others can be brutal.
[...] In theory, the New York Times lawsuit could end in a similar way. The Times claims OpenAI used millions of Times articles to train GPT-4. At $25,000 per article, OpenAI and Microsoft could owe billions of dollars to the New York Times alone.
But there is an important way that OpenAI differs from MP3.com. It already has tens of millions of users and passionate defenders ready to explain how its products help them be more creative and productive.
A ruling that shut down ChatGPT completely would disrupt the lives of its users and could undermine America’s leading position in AI technology. Judges do not like to make waves and will be reluctant to do that.
So, judges could be motivated to view OpenAI’s legal arguments more sympathetically—especially the argument that generative AI is a transformative use. And even if they do find that AI training is copyright infringement, they have off-ramps to avoid putting AI companies out of business. Courts might award much smaller damages, or they might press the parties to settle the lawsuit before reaching a final decision.
[...] Generative AI developers have some strong arguments in response to copyright lawsuits. They can point to the value that their AI systems provide to users, to the creative ways that generative AI builds on and remixes existing works, and to their ongoing efforts to reduce memorization.
But all of these good arguments have something in common: they take copyright issues seriously. These responses acknowledge that generative AI is built on a foundation of training data, much of which is copyrighted, and then try to show that all of this copying is justified rather than that it is irrelevant... (MORE - missing details)
EXCERPTS: . . . So far we’ve focused on the legal vulnerabilities of AI companies. But the defendants could have one big factor working in their favor: judges reluctant to shut down an innovative and useful service with tens of millions of users.
Judicial opinions are framed as technical legal analysis, but they are often driven by a judge’s gut feeling about whether a defendant has created a useful service or is merely seeking to profit from the creative efforts of others.
[...] A judge who feels a product is valuable will look for ways to allow it to stay in business. On the other hand, a judge who feels a product unfairly profits from the creative work of others can be brutal.
[...] In theory, the New York Times lawsuit could end in a similar way. The Times claims OpenAI used millions of Times articles to train GPT-4. At $25,000 per article, OpenAI and Microsoft could owe billions of dollars to the New York Times alone.
But there is an important way that OpenAI differs from MP3.com. It already has tens of millions of users and passionate defenders ready to explain how its products help them be more creative and productive.
A ruling that shut down ChatGPT completely would disrupt the lives of its users and could undermine America’s leading position in AI technology. Judges do not like to make waves and will be reluctant to do that.
So, judges could be motivated to view OpenAI’s legal arguments more sympathetically—especially the argument that generative AI is a transformative use. And even if they do find that AI training is copyright infringement, they have off-ramps to avoid putting AI companies out of business. Courts might award much smaller damages, or they might press the parties to settle the lawsuit before reaching a final decision.
[...] Generative AI developers have some strong arguments in response to copyright lawsuits. They can point to the value that their AI systems provide to users, to the creative ways that generative AI builds on and remixes existing works, and to their ongoing efforts to reduce memorization.
But all of these good arguments have something in common: they take copyright issues seriously. These responses acknowledge that generative AI is built on a foundation of training data, much of which is copyrighted, and then try to show that all of this copying is justified rather than that it is irrelevant... (MORE - missing details)