[LIS-Forum] Should AI companies take free advantage of public domain material gathered by a large community of 'open' folks over the years?