[kianlocke]

In my view it's categorically wrong to assume that these systems will cost a billion dollars forever, and if we write the laws without that understanding, we are only helping billion-dollar corporations.

In cases where these systems were trained on data that was literally stolen from closed, user-gated sites, this is one thing -- for those trained on data that was pulled from the open web, there are a number of cases relating to scraping sites that publicly make available information, and they tend to fall in favor of the scraper (see LinkedIn for a prime example).

[mvdtnz]

Whether the system cost a billion dollars or not isn't the point. The point is that computer programs are categorically different to human brains because of the scale and perfect reproducibility.