The Coming IP Fights That Could Reshape AI

The next major battle over artificial intelligence may not be about what AI can do. It may be about what it can use, and who owns what it creates.

Reuters reports that OpenAI, Microsoft, The New York Times and a group of authors are asking a federal court to address whether copyrighted books and news articles can be used to train AI models under the doctrine of fair use.

But that's only one side of a much larger intellectual-property question.

On the input side, courts are beginning to grapple with whether AI companies can train their models on copyrighted material without permission.

On the output side, an equally important question is developing: Can inventions, content and other valuable assets created with AI actually receive intellectual-property protection?

Current U.S. law still places considerable importance on human involvement. The USPTO maintains that only natural persons can be inventors, although AI-assisted inventions can qualify for patent protection. The Copyright Office similarly distinguishes between AI used as a tool by a human creator and material generated by AI without sufficient human authorship.

That creates a fascinating tension.

Companies are investing enormous amounts of money in AI partly because of what it can create. Yet the more independently AI creates, the more complicated the question may become of who owns, or can protect, the resulting asset.

For businesses, this is much bigger than copyright litigation. It touches patents, copyrights, licensing, data rights, contracts and ultimately the value of AI-generated intellectual property.

The first cases are beginning to work their way through the courts and agencies.

The larger fight over what AI can use, what AI can create, and who owns the value on either side may just be beginning.

-Matthew Murphy