The US Copyright Office's draft report asserts that training AI on copyrighted works for commercial use likely violates fair use laws. While utilizing large language models for research may be permissible, the report underscores that commercial applicationsâparticularly those competing with original materialsâoverstep legal boundaries. The findings challenge the positions of leading AI firms, which argue that any published content should be free to use. The dismissal of the agencyâs head, Shira Perlmutter, shortly after the report's release raises concerns about the influence of these companies and their lobbying efforts against copyright regulations.
The report concludes that while using LLMs based on copyrighted data for research may be acceptable, commercial applications that compete with original works likely breach fair use laws.
AI companies have generally operated under the belief that all published content should be accessible without compensation, but the Copyright Office report contradicts this view.
Collection
[
|
...
]