Defeat in a recent libel case in Italy has given Google a headache over the output of its 'autocomplete' function, Google Instant. Google has also recently published the latest figures for its Transparency Report giving details of the volume and type of requests for information it has received from governments in the latter six months of 2010.
Google and autocomplete
Autocomplete suggests additional words to users in real-time, related to the search terms they enter. The service has sparked controversy because of the appearance of suggestions to accompany users' search terms that could be seen as derogatory or offensive.
Following the recent case, Google will now have to take responsibility in Italy as a content provider for the suggestions that are offered, including where these are defamatory. This rules out use of the 'safe harbour' defence for internet service providers acting as "mere conduits" under the E-Commerce Directive.
The case involved an Italian entrepreneur who brought a libel action against Google after the search engine's autocomplete function linked his name to the Italian terms for 'conman' and 'fraud'. These were, it was claimed, detrimental to the public image of the claimant "as an entrepeneur and provider of educational services in the field of personal finance". The Court of Milan ruled in favour of the claimant, finding that Google was responsible for the defamatory nature of the terms.
The court rejected Google's claims that it was not responsible for the content of autocomplete suggestions, on the basis that these were "predicted by computer algorithms based on searches from previous users, not Google itself". Under the E-Commerce Directive, providers are exempt from liability for content they allow access to but which they did not create. Service providers' actions must be "of a mere technical, automatic and passive nature"; in this case, Google was held to have produced the content in question, albeit through automated means.
It remains to be seen what the ruling means for the future of Google Instant in Italy. Google currently operates a narrow removal policy restricted to censorship of "pornography, violence, hate speech, and terms that are frequently used to find content that infringes copyrights". The user-driven nature of the service means that significant challenges are likely to exist around the practicalities of removing libellous search terms.
Google has recently made available the latest figures for its Transparency Report. The figures published at the end of June cover the period from July to December 2010. In publishing these figures Google has offered more data and has also updated the look of the report.
The data is broken down on a country by country basis. Top of the list of those governments making the most requests for user data is the United States, with Brazil second and the UK in fourth. In addition the United Kingdom made 38 requests for removal of content from Google services. The main case highlighted involved a request to remove fraudulent ads that linked to scams. As a result of this request over 93,000 items were removed. The report also identifies the reason for removing the content and gives a percentage figure for the number of requests that are complied with.
Launching the latest figures the Official Google Blog noted that the aim of publishing the figures is to show “how the web is shaped by government influence and how Google responds to requests for information and removals”. It also gives interesting insight into the actions being taken by authorities in respect of this type of data.