Artificial intelligence and libel
In the wake of the multibillion-dollar settlement between state attorneys general and Meta ( Facebook / Instagram) over addiction by young users, the question of liability for harms caused by big tech companies is hardly moot.
Libel is one of those areas of product liability. It’s a new problem with a new technology, but even if expression by an AI bot is protected by the First Amendment or Section 230 of the Communications Decency Act, it may still be liable for damaging falsehoods.
On Sept. 14, 2026 a case brought against Google for AI libel was allowed to go forward in an Illinois court. Google argued that it used a disclaimer to say that AI Overviews are merely “starting points” and not statements of fact. Google also argued that it was simply linking to information. But Google’s defenses were not sufficient to prevent the case from going forward, the judge said in Keene v Google.
According to a New York Times article “The AI Industry’s New Worry” Sept 18, 2026, OpenAI and Anthropic could protect themselves from liability risk if they could prove they took reasonable care to protect against foreseeable risks. Those measures could include rigorous model monitoring and other guardrails, as well as slowing the release of new models.
What would that entail? The Journal of Free Speech Law has devoted a lot of ink to the issue, especially Vol.3 No. 2, 2023. The most comprehensive 0verview is by Eugene Volokh: “Large Libel Models”
“… When the “actual malice” standard applies, the standard might be satisfied if an AI company has received actual notice of particular spurious information being produced by its software but has refused to act. This would in practice require such companies to implement a “notice-and-blocking” system, loosely similar to “notice-and-takedown” systems required under the DMCA as to copyright and trademark infringements. (There is also) the possibility of negligence liability, when such liability is authorized under libel law, by analogy to negligent design product liability.”
Volokh also argues that Section 230 of the CDA does not indemnify AI companies from harms caused by false information / hallucinations.
Current AI libel cases
Wolf River Electric v Google — A solar contractor in Minnesota noticed a sudden drop-off in business and realized that apparently false information was being spread through Google’s artificial intelligence chatbot. This is one of a half dozen cases that will attempt to define content that was not created by human beings as defamatory.
Walters v OpenAI — Nov. 21 — A Georgia state court granted a motion to dismiss a case brought by radio broadcaster Mark Walters against OpenAI, an artificial intelligence developer of ChatGPT. Walters alleged that ChatGPT had libeled him with false information about supposed financial fraud. In its decision, the court noted that the service uses a disclaimer: “ChatGPT may produce inaccurate information about people, places, or facts.” Walters also testified that he was not harmed by ChatGPT and that he had not contacted the service and asked it to withdraw the false information.
(MORE TO COME HERE)
MORE
New York Times, Nov. 12, 2025 , Who pays when AI is wrong?