Baidu is, broadly speaking, the Chinese equivalent of Google. Last week, a Chinese district court found it liable for copyright violation. The act of infringement? Automatically constructing and maintaining an index of links, some of which were to infringing files.
The Court ordered Baidu to pay damages of 68 000 yuan (AUD$8 400). The company now faces a litany of followup suits, including actions by Universal Music, EMI and Sony BMG. Baidu’s lawyers remain confident, however:
‘We believe that the district court order was based on a misunderstanding of the search engine technology and therefore is without merit’, Decheng Li, a lawyer with Zhonglun WD Law Firm, which represents Baidu, said in a statement. …
Baidu has maintained it is not at fault as it merely provides the search technology and not the offending downloading service.
The plaintiff in this case was Shanghai Busheng (‘Push Song’), a Chinese music publisher. It argued that because Baidu’s search technology provided links to websites which allowed users to illegally download MP3s subject to copyright, it should be liable for infringement.
The acceptance of this argument is problematic for several reasons. First, if liability is deemed to turn on a hyperlink, linking — the fundamental birthright of hypertext — would become a risky business. Ultimately, it may cause the web to lose its interconnectedness, and hence its utility.
Second, how far does liability extend? If I link you to an MP3 search engine using which copyrighted works can be found and illegally downloaded, but don’t provide direct links to the MP3 files (like Baidu did), am I liable for your subsequent infringement? What if I link to the results page of a specific query? What about a Google search for “MP3 search engine”? What if I just link to the software or technology by which such infringement is made possible, like KaZaa or BitTorrent? Firefox? The basis for second or third degree liability is even more tenuous, and the consequences far more devastating.
Third, it is a form of secondary infringement unknown to other civil wrongs. Linking to defamatory content, without more, cannot (to my knowledge) give rise to liability. Linking to an article certainly isn’t copyright infringement of that article. Linking to information which, if acted upon, would be illegal, is not — at least in the absence of a supporting procurement offence — itself illegal. Holding website owners liable for the actions of users upon the content of still other users, which they may neither endorse nor realise exists, is thus unprecedented.
Fourth, links change. Today’s cheapmovies.com.au, a legitimate DVD retailer, may be tomorrow’s BitTorrent movie download centre. There are thus significant evidentiary problems associated with accusing an automated content indexer of infringement when the content — and perhaps even the associated proprietary rights — to which they are alleged to have linked did not even exist at the time the hyperlink was created.
Fifth, there is no inducement or authorisation because automated content aggregators are among the most neutral parties on the internet. Indexing is simply a recursive algorithm, and all raw content is stored regardless of merit or legality. Correspondingly, search engines promote themselves as information resources, not piracy hubs. To filter by keyword would be antithetical to the notion of a keyword-based search engine. Think of the chilling effect upon informational goods if Google and Baidu were responsible for the content of the 8 billion or so pages in their respective indexes, all of whose owners hold some form of intellectual property in their work.
These concerns are similar to those articulated in response to the MP3s4free.net and Outlaws.com cases, but are significantly more pressing where, as here, the liable party is a neutral indexer of content and not actively encouraging users to infringe copyright. In light of Grokster and Sharman License Holdings, it is unlikely that a party could be held liable for linking without something more (Moorhouse would suggest ‘sanctioned, approved or countenanced’). For Baidu, however, whose stock price tumbled 28 per cent last week, rulings like these threaten to destroy the commercial viability of automated search indexing.
