A recent German ruling held Google liable for errors introduced in AI-generated overviews provided by the company. Security experts Bruce Schneier and Nathan Sanders have publicly commented on the decision, arguing in favor of reinforcing corporate responsibility for such mistakes.
The argument: AI agents as agents of the deployer
Bruce Schneier and Nathan Sanders contend that AI agents should be treated legally as agents of the person or organization that deploys them. They note that if a company hired human writers to produce summaries, the company would be responsible for any inaccuracies; the same principle, they argue, should apply when the summaries are produced by AI.
Why this matters
According to Schneier and Sanders, allowing businesses to hide behind the excuse of "faulty AI" would amount to a substantial concession to companies and create harmful incentives. It could encourage replacing human professionals—writers, lawyers, doctors—with cheaper AI systems that firms could use to avoid liability when mistakes occur.
Conclusion
The German ruling and the Schneier–Sanders argument highlight a pressing legal and policy question: how should legal systems allocate responsibility for harms caused by deployed AI agents? Recognizing corporate liability for AI errors could deter perverse business incentives to substitute human expertise in order to evade accountability.
