In late 2023, the Times brought a copyright action against Microsoft and OpenAI for impermissibly copying Times material to “train” the generative AI behind the defendants' AI tools (including ChatGPT). The Times complaint alleges that the AI models do not merely ingest and transform copyrighted material, but that their output sometimes consists of straight copies of large passages of Times content.
Why It Matters
Copyright owners have the right to prevent appropriation of their material, but a long-held concept allows users to “transform” copyrighted material under certain circumstances. So far, AI operators have argued that their engines “transform” the copyrighted materials that are ingested when training the AI engine. If the Times has accurately identified flat-out copying, it will ignite a scramble in the AI world to figure out how to filter results or otherwise tweak the output of AI engines to avoid plagiarism. This may impose costs and other burdens on AI operators. It is not just news organizations that have the right to protect their content, however. The rulings that come from the Times and other cases will shape how AI tools can treat content from and about each of us in the years to come.
Subscribe to Taylor English Insights by topic here.