A German court has delivered a significant blow to Google's ability to distance itself from the content of its AI-generated search summaries. The ruling suggests that these digital overviews are not simply neutral aggregations of online data, but rather an active reworking of information by the tech giant – and as such, Google must take responsibility for their accuracy. This decision marks a turning point in the ongoing debate about accountability for AI-generated content.
The German court's verdict hinges on a fundamental distinction between 'carriers' and 'publishers'. Traditionally, phone companies are considered carriers, transmitting data without liability for its content. Newspapers, by contrast, are publishers – exercising editorial control and thus accountable for the accuracy and legality of their published words. Internet companies like Google have often navigated this divide, claiming carrier status to avoid liability while exercising publisher-like control over content when beneficial.
However, the German court's reasoning places AI-generated summaries firmly in the 'publisher' camp. This interpretation acknowledges that these digital overviews involve a level of editorial discretion – akin to a human editor crafting a newspaper article or an original essay. By doing so, the court implies a greater duty of care and responsibility for their accuracy.
This ruling is not isolated. Two years ago, Air Canada was held liable after its AI chatbot offered a customer a discount the airline later refused to honour. The court ruled that the airline was as responsible for its chatbot's statements as it would be for information on its official website. This precedent, alongside the German decision, indicates a growing trend towards corporations being held accountable for the actions and statements of their deployed AI systems – treating them as agents of the organisation.
For UK businesses and consumers, this evolving legal landscape has significant implications. Companies deploying AI chatbots, summarisation tools or other generative AI in customer-facing roles may face increased scrutiny regarding the accuracy and reliability of the information provided. Consumers, in turn, may gain stronger legal grounds to seek redress if misled or harmed by faulty AI outputs. Experts like Bruce Schneier and Nathan E Sanders argue that such liability is crucial for demanding accuracy and holding companies accountable – preventing businesses from using 'faulty AI' as an excuse for harmful or misleading information.