Lawsuits against artificial intelligence (AI) companies like OpenAI and Microsoft highlight an uncomfortable truth for these tech giants. Namely, under current copyright law, these companies are very likely fighting a losing battle. The reality is that training AI models requires utilizing vast amounts of copyrighted material, and our existing intellectual property framework is purpose built to prevent and punish this practice. But we shouldn’t accept this legal status quo. Congress must act quickly and decisively to create a clear copyright exemption for AI training. The stakes are simply too high to let 20th century copyright law strangle 21st century innovation.
The fair use doctrine, which AI companies lean heavily on in their defense, wasn’t designed with machine learning in mind. This should come as little surprise, as the general contours of existing copyright law date back to the Copyright Act of 1976. This law was written when cutting-edge technology meant pocket calculators and eight-track tapes. While Congress …