In the decision of 14 May 2013 (Docket no. VI ZR 269/12) the Federal Court of Justice (“FCJ”) held that predictions in Google’s autocomplete function could lead to an infringement of personality rights and secondary liability of the search engine operator. FCJ lifted the prior decision of the Appellate Court of Cologne of 10 May 2012 (Docket no. 15 U 199/11) that took a different view and rejected the claims of the plaintiff by denying any personality right impairment a priori.
Predictions – Statement of the search engine operator?
In the case at issue, the plaintiff claimed infringement of his personality right when his name was associated with the terms “fraud” and “scientology” by autocomplete predictions. Predictions made by autocomplete are a reflection of the search activity of all web users and consider e.g. keywords frequently used by others. They are guided by algorithms that take different parameters into account.
The Appellate Court did not consider the predictions as being a statement of the internet service provider (“ISP”) and outright refused negative connotations and infringement of personality right of the plaintiff. According to the Appellate Court the average user understands that predictions are the outcome of an algorithm based software process without implying any personal statement of the ISP.
Federal Court of Justice considers secondary liability
In contrast, the FCJ took a different view on the expectation of the average user of search engine services and held that the majority of users would assume that only reasonable and content-related predictions are associated with the search terms.
On these grounds, FCJ deduced a negative nexus between the predictions at issue (“fraud”, “scientology”) and the plaintiff’s name, entered as a search keyword, resulting in a personality right infringement in the end. According to FCJ, this infringement of a personality right shall be directly attributed to Google as the ISP has developed and used the autocomplete system by evaluating user behaviour and providing respective proposals for optimizing search results.
At least the Court did not assume liability without limitations and refused a general obligation on ISPs to actively monitor predictions for infringements of personality rights. Instead, secondary liability as a (so-called) “interferer” (German: “Störer”) requires, as a first step, prior notification of an alleged infringement and, as a second step, disregard of an obligation to undertake reasonable measures to prevent prospective infringements of the same kind by the ISP.
Two aspects of the Court’s reasoning raise doubts and objection. First, the decision of FCJ appears to be internally inconsistent. Although the Court considers the predictions made by autocomplete to be Google’s own content, the principles of secondary liability shall apply. This approach does not harmonize with existing case law. Second, and this is the more important aspect, the core assumptions of FCJ do not properly reflect the market standards and conventions and the view of internet users. At least, ongoing discussion on the algorithm mechanism will have the effect that average users will expect less and less the predictions to be a selective suggestion, rather than (as it is) a technical result of a software process. Google’s approach to provide additional information on autocomplete function through a link in the predictions window will, of course, strengthen this development.
Inconsistency of European case-law
For the time being, legal assessment on Google’s autocomplete remains inconsistent in European case law. In contrast to the FCJ, other courts rejected similar claims raised by the plaintiffs (cf. District Court of Jura, decision of 4 February 2011, docket no. CC 117/2010, Switzerland; Tribunale di Pinerolo, decision of 30 April 2012, docket no. 696/2012, Italy.) Harmonisation of European case law on autocomplete function will, most likely, not be achieved until a decision of the European Court of Justice.