A German court has ruled that Google is legally responsible for the content generated by its AI search summaries, marking a significant development in the legal treatment of artificial intelligence. The decision, handed down earlier this month, rejected Google's defenses that users could verify information themselves or that people generally know AI-generated content should not be blindly trusted. The court held that the AI summaries are reflections of the company and 'above all an expression of Google's business activities.'
Carriers vs. Publishers: A Historic Distinction
The ruling is the latest skirmish in a decades-old battle over internet publishing. Historically, information distributors fell into two categories: carriers and publishers. A phone company, as a carrier, transmits whatever users say without liability for the content. A newspaper, as a publisher, decides what to publish and is liable for defamatory or illegal material. Internet companies have long tried to play both roles, claiming carrier status when convenient and publisher status when advantageous. Section 230 of the 1996 Communication Decency Act enshrined this straddling by shielding internet providers from liability for the speech of others on their platforms.
For years, debate has centered on how to apply this law to social media platforms. Early platforms displayed posts in reverse-chronological order, behaving like carriers. But newer platforms like Facebook curate feeds with algorithms, acting more like publishers by making editorial decisions. Some experts argue Section 230 has gone too far and needs reform, while others see it as essential to the modern internet.
AI Overviews: A New Category of Liability
Google's AI overviews differ fundamentally from traditional search, which courts have treated as archiving and facilitating access to third-party editorial content. AI overviews do not simply quote and republish words from websites; they rewrite others' words, exercising editorial discretion akin to a newspaper article or an original essay. This distinction was central to the German court's reasoning.
The ruling extends beyond Google. Any company deploying AI to generate summaries or content could face similar liability. For instance, a restaurant review site with AI summaries or a publisher using AI to summarize its own material could be held accountable for inaccuracies. Accuracy matters, and liability is a key mechanism for the public to demand accuracy and hold companies accountable when they cause harm.
Precedents: Air Canada and the Duty of Care
Two years ago, Air Canada learned a similar lesson. Its AI chatbot promised a discount the company later rescinded. In court, Air Canada argued it was not responsible because the chatbot was a 'separate legal entity.' The court sided with the flyer, ruling the airline was as responsible for its chatbot's statements as for its website content. This precedent suggests corporations have a duty of care for the performance of the AI chatbots they employ.
As Bruce Schneier and Nathan E Sanders argue, 'AI agents are agents of the person or organization that deploys them – and should be treated by the law as such.' If a company hires human writers to produce summaries, it is liable for inaccuracies. If a human agent signs contracts, the company is bound by those contracts. Allowing businesses to hide behind faulty AI would be a massive handout and introduce disastrous incentives for corporate misbehavior.
Implications for AI Assistants and Commerce
We are rapidly moving toward a world where AI-powered chatbots will be at the other end of many corporate communications channels. It makes no sense for a company to honor its statements when it wants to and disavow them when it does not. Visa and OpenAI recently announced a partnership to build personal AI agents that can make purchases on behalf of users. Questions arise: Will Visa take responsibility if its AI makes an unwanted purchase? Properly allocating liability is key to making such systems trustworthy.
If the German ruling holds, it could be devastating for Google's AI Overview feature. Tests from earlier this year found that Google's AI summaries had mistakes about 10 percent of the time. With more than 5 trillion searches per year, that translates to approximately 16,000 erroneous summaries every second. While most errors are benign, some cause harm, are defamatory, or trigger liability.
Ongoing Cases and Outcomes
Earlier this year, Google's AI summary falsely identified Canadian fiddler Ashley MacIsaac as a sex offender. His lawsuit, filed in Ontario, is ongoing. If Google is forced to invest in improving its AI system until such errors are exceedingly rare, that would be a positive outcome for users and subjects of search like MacIsaac.
More generally, liability concerns could mean that many current use cases for AI agents may not be commercially viable. Companies may not profitably operate AI lawyers, doctors, or media influencers if held responsible for their outputs. As Schneier and Sanders conclude, 'There's nothing in the law that requires us to accommodate AI systems if they are fundamentally untrustworthy, just as we don't need to accommodate untrustworthy human systems. Any company that won't stand by the statements its agents make – whether human or AI – doesn't deserve users' time or money.'
