WikiHow Sues OpenAI: The Creative Commons Loophole Closes

WikiHow has sued OpenAI for using its content without permission in AI training. What makes this case legally distinct from *The New York Times* or Getty Images suits is that WikiHow published its content under Creative Commons licensing—free to use, but not for commercial purposes. That NC clause may be the cleanest copyright argument against training data use yet assembled.

In plain terms

WikiHow publishes free step-by-step guides for doing almost anything. It made this content free on purpose, for public use. WikiHow is now suing OpenAI for using that content to train a commercial AI product and reproduce its substance in ChatGPT—without authorization or compensation.

The lawsuit

WikiHow's case, reported by Medianama, alleges that OpenAI scraped and copied WikiHow's copyrighted articles to train its models, use them in RAG systems, and reproduce their substance in ChatGPT responses. The claim is copyright infringement, with the added dimension that WikiHow's content was published for public benefit—not for extraction by a commercial AI product.

This adds WikiHow to the growing list of content creators in legal conflict with AI labs: The New York Times, Getty Images, individual authors through the Authors Guild, music publishers. The suits pursue the same fundamental question through different plaintiffs: does training a commercial AI on copyrighted content constitute infringement?

Why WikiHow's case is worth watching

WikiHow's content is explicitly instructional and procedural—step-by-step guides that AI models can reproduce almost verbatim in response to user questions. The RAG allegation (using WikiHow articles in retrieval-augmented responses) is potentially cleaner than the training data allegation, because the reproduction is more direct and the commercial harm to WikiHow—users who would otherwise visit WikiHow get their answer from ChatGPT—is more traceable.

The settlement economics

Labs that trained on everything and worried about licensing later are carrying increasing legal exposure. The question is whether eventual settlement cost—likely paid through licensing deals rather than court judgments—is a rounding error or a structural operating cost. As more plaintiffs file, the negotiating position of content owners strengthens.

So what

If you're building AI products and your training or retrieval data includes content from sites with their own licensing terms, WikiHow's suit is the moment to review data provenance and usage claims. The "we didn't know" and "it was freely available" defenses have been shrinking for two years. This case adds "it was freely available but for non-commercial public benefit" to the challenge list.