Does scraping data for generative AI training infringe copyright?

[El Presidente]

Does scraping data for generative AI training infringe copyright?

We are about to find out with heavy hitters heading to court. 

https://www.nytimes.com/2023/12/27/business/media/new-york-times-open-ai-microsoft-lawsuit.html

Complicated and interesting area. 

In obtaining content, is  Open AI really that different to Napster?

[MrBirdman]

In short, probably yes.

Will it be protected under fair use? Much harder to say. Courts will be aware that excluding generative AI from fair use could hurt the industry, but there is a solution: pay for access. News organizations are already changing their TOS to prohibit using their material for training, and that’s a contractual agreement, so AI companies will need to confront the freeriding issue one way or the other. 

It’s a very difficult issue though.