Background
Munich Regional Court I has ruled that Google may not further disseminate false statements from an “AI Overview”. In the underlying case, Google’s AI warned users about a Munich-based publisher, described it as known for dubious business practices and created links to an alleged fraud scheme. Subscription traps, debt collection claims and the recommendation to seek legal assistance were also mentioned.
The case was brought by the affected publishing house and its subsidiary. The trigger was search queries in which users entered the publisher’s name together with the word “fraud scheme”. According to the claimants, the AI processed information about other companies and used it to create false connections that impaired their reputation in business dealings.
The court’s decision
Google argued that the data came from third-party websites. The search engine merely summarised this content automatically; users could click on the sources and check them themselves. Munich Regional Court I did not follow this argument.
In the court’s view, the “AI Overview” differs fundamentally from a traditional list of search results. Google does not merely display links or short snippets. The AI processes information, restructures it and formulates a new, independent answer. According to the report, individual claims made by the AI also did not appear in the linked sources.
Corporate personality rights and AI content
The claimants based their injunction claim on corporate personality rights. These rights protect companies from damage to their reputation in commercial life caused by false allegations. The court therefore had to weigh the company’s protection against freedom of expression.
The court classified the statement about the “fraud scheme” as an expression of opinion because evaluative terms such as “dubious” and “fraud scheme” dominated. This did not help Google, however. Even an expression of opinion loses considerable weight in the balancing exercise if it is based on demonstrably false facts. According to the report, the AI relied on links to subscription traps that did not actually exist.
Significance for businesses
The decision is relevant for companies, brands, media houses and platform operators. It shows that AI-generated search responses are not automatically treated like neutral lists of links. Where information is summarised, evaluated and presented as an independent answer, it may become legally attributable content.
For affected companies, the decision may mean that reputation-damaging AI outputs do not have to be dismissed merely as a technical error. However, the specific content, the provability of its inaccuracy and the legal assessment in the individual case remain decisive.
To the point
- Munich Regional Court I prohibited Google from further disseminating the challenged AI statements.
- The “AI Overview” was not treated as a mere list of links.
- The court regarded the AI response as Google’s own statement.
- Even evaluative statements may be unlawful if they are based on false facts.
- Companies should carefully document reputation-damaging AI outputs.
Source: Bayerische Staatskanzlei