Writing in NLJ this week, Hanna Basha, partner, and Jamie Hurworth, legal director, in Payne Hicks Beach’s dispute resolution team, say anyone who incorporates AI-generated text into a report or post will generally be treated as adopting it, regardless of whether they knew it was false.
AI providers may face a more novel claim. Unlike traditional hosts, they generate the words through systems they design and operate. A German court has already held Google responsible for defamatory AI Overview content, although its decision is not binding here. A provider’s position may become more exposed once it is notified of a recurring false allegation and can prevent its repetition.
The authors caution that disclaimers will not necessarily help; serious harm, meaning, honest opinion and public-interest defences all pose difficult questions.
