The owner of the Daily Mail is gearing up for a legal battle with Google over claims the tech giant used hundreds of thousands of online news stories to train its ChatGPT rival without permission.
Daily Mail and General Trust (DMGT), the publishing group controlled by Lord Rothermere, is understood to have sought legal advice as it considers potential action.
DeepMind, Google’s artificial intelligence (AI) division, allegedly harvested a vast cache of around 1 million news articles from the Daily Mail and CNN websites to help develop its chatbot, Bard.
It is claimed that the tech giant targeted them because both use bullet points to summarise key points before the main text of a story. Google allegedly used the articles to test Bard’s capabilities by removing words from the bullet points and asking the AI to fill in the gaps based on the rest of the story.
However, it allegedly used the articles from the Daily Mail and CNN websites without either copyright holder’s knowledge or permission. Around three-quarters of the articles used in the dataset are believed to have come from the Mail, while the remainder were taken from CNN’s website.