First, this is about an AI being trained on a competitor's database. It's not about training an AI on everything that's on the web, much of which is copyrighted. That copyright issue has not yet been litigated, so far as I know. Still, it would be hard to argue that training on copyrighted material is fine, when the issue here is training on something that is copyrighted.
Second, this is a circuit court decision. This is going to have to go through an appeal to the district court, and then the supreme court, before it becomes settled law.
Third, it's a Third District decision. It doesn't apply in California (or, in fact, in most of the country).