Are autonomous AI overviews legally binding in contract disputes?

"Are autonomous AI overviews legally binding in contract disputes?" is really two separate questions, and conflating them is where the risk hides. The first is whether an AI-generated summary itself creates or alters a legal obligation — generally, no, it doesn't have that status any more than a marketing brochure does. The second, more consequential question is whether the company or lawyer who relied on that AI output is on the hook when it turns out to be wrong. On that question, there's now a real, if still developing, body of case law, and the answer is consistently: yes, you're responsible for what your AI told someone.

The precedent: a company is liable for what its chatbot says

The clearest example is Moffatt v. Air Canada, decided by the British Columbia Civil Resolution Tribunal in February 2024. A customer asked Air Canada's website chatbot about bereavement fares, and the chatbot gave him incorrect information about how to retroactively claim the discount. When he tried to use it, Air Canada refused and argued in the tribunal that the chatbot was "a separate legal entity that is responsible for its own actions." The tribunal rejected that outright, ruling Air Canada liable for negligent misrepresentation and ordering it to pay damages. The reasoning matters more than the dollar amount: a company doesn't get to disclaim responsibility for its own customer-facing AI tool just because a human didn't type the specific words.

That principle — you own what your deployed AI tells people — is the practical answer to "is it legally binding." The AI's output isn't itself a binding contract clause, but the business that put the AI in front of customers can absolutely be bound by ordinary doctrines like negligent misrepresentation, the same as if an employee had said it.

The other side: courts sanctioning lawyers for trusting AI citations

A separate but related problem has emerged specifically in contract and litigation disputes: lawyers using AI tools for legal research and getting fabricated case citations in return, then filing them without checking. This is no longer rare. A database maintained by legal researcher Damien Charlotin has tracked well over a thousand instances of courts flagging AI-hallucinated content in filings. Consequences have escalated: a federal judge in Oregon sanctioned two lawyers a combined $110,000 for submitting 23 fabricated citations and eight invented quotations, the Sixth Circuit sanctioned attorneys for over 24 fake citations in Whiting v. City of Athens, and the Ninth Circuit suspended two attorneys from practicing before it for six months over similar conduct.

This matters directly for contract disputes because legal research tools marketed specifically to lawyers aren't immune. A 2024 Stanford study (updated and reconfirmed in 2025) tested Thomson Reuters' Westlaw AI-Assisted Research and LexisNexis's Lexis+ AI against real legal queries and found hallucination rates of 33% and 17% respectively — even though both are retrieval-augmented tools built specifically to reduce this problem. The lesson for anyone using AI to research a contract dispute: treat every case citation an AI gives you as an unverified lead, not a fact, until you've pulled up the actual case yourself.

What this means practically

If you're a business deploying an AI tool that answers customer questions about contracts, policies, or terms: assume you are legally responsible for what it says, the same as a human representative, and budget for review/correction processes accordingly — don't treat "the AI said it" as a liability shield. If you're a lawyer or paralegal using AI for contract research: verify every citation against the primary source before it goes in a filing. Courts have made clear that "the AI hallucinated" is not a defense; the duty to verify sits with the human who signs the document.