Who Owns an AI-Generated Listing Description? The Copyright Question Is Still Open.
When a real estate agent uses ChatGPT to write a listing description and that description ends up in the MLS, who owns it? The agent? The brokerage? OpenAI? Nobody? This question sounds academic until someone tries to reproduce, scrape, or repurpose that content — and then it matters quite a bit.
The Copyright Office's Current Position
The U.S. Copyright Office has issued guidance indicating that purely AI-generated content — content produced without meaningful human creative input — is not eligible for copyright protection. Copyright requires human authorship. A work created entirely by prompting an AI system, without the human exercising meaningful creative control over the expression, falls outside the scope of what copyright protects.
The practical implication: a listing description generated by dropping a few property details into ChatGPT and accepting the output verbatim may not be copyrightable by anyone. It sits in a legal gray zone — not clearly in the public domain, but not clearly protected either.
Where Human Input Changes the Analysis
The analysis shifts when a human exercises meaningful creative control over the output. Editing the AI's draft, selecting from multiple versions, combining AI-generated and human-written passages, or providing detailed creative direction that shapes the expression — these are forms of human authorship that can support a copyright claim in the resulting work.
This creates an interesting practical implication: the more heavily an agent edits and shapes AI-generated content, the stronger their copyright claim. The agent who treats AI as a first-draft tool and then rewrites substantially has a better claim than the one who copies and pastes.
The MLS Data Layer
There's a separate and somewhat underappreciated issue around MLS data rights. Most MLS agreements include provisions about who owns listing data and under what circumstances it can be reproduced or licensed. As AI tools increasingly rely on scraping and training on MLS data — and as AI-generated content flows back into MLS systems — the contractual frameworks that govern that data are being tested.
Several major MLS systems have updated their terms of service to address AI training and data scraping, generally prohibiting unauthorized use of listing data to train AI models. Whether those provisions are enforceable in all contexts remains to be litigated.
The Practical Takeaway
For agents and brokerages today: don't assume AI-generated marketing content is protected just because you paid for the tool that generated it. If you want to protect your content — for example, if you're building a brand around distinctive listing copy — make sure you're exercising genuine creative input, not just prompting and publishing.
The legal framework around AI and intellectual property is moving faster than most people realize. The real estate industry, which generates enormous amounts of text and image content, is going to find itself at the center of several of these emerging disputes.
- Jason