A federal district judge in New York City issued a troubling ruling today holding that an electronic news clipping service infringed copyright when it republished excerpts of news stories in search results for its clients seeking news coverage based on particular keywords.

The case is Associated Press v. Meltwater. Meltwater is a private subscription service that scans news sites for stories relevant to its clients and then delivers the search results in the form of short excerpts from, and links to, the original articles. News service Associated Press claimed the search results infringed its copyrights in the news articles included within them. Meltwater argued that its service was a noninfringing fair use. EFF filed an amicus brief supporting Meltwater.  

The court’s fair use analysis is worrisome in at least three respects. First, the court concluded that Meltwater’s purpose was not transformative because it neither added commentary or insight to the excerpts it sent to customers nor served the same information-finding goal as a search engine like Google.  Why not? In a nutshell, apparently because it was not public and not very successful at getting its customers to click through to the original articles.  Given that the court devoted several pages of text explaining how Meltwater was like a search engine (i.e., it scans the web for news and creates an index that allows customers to search for relevant information) it is difficult for us to see why the fairness of its purpose should turn on its success at getting customers to actually click through the links it provided. The court took great care to distinguish what it saw as “legitimate” search engines – but we can expect more litigation over what counts as “legitimate.”

Second, the court implicitly adopted AP’s dangerous “heart of the work” theory.  AP contended that sharing excerpts of a news article must weigh against fair use if those excerpts contain the lede.  The court stressed that the lede is “consistently important” and takes “significant journalistic skill to craft.” But that is beside the point – there is no extra protection because something is extra difficult. More important to the fair use analysis is the fact that (1) is primarily factual; and (2) contains precisely the information the user wishes to make known to others. As we explained in our amicus brief, this case illustrates why the heart of the work doctrine does not mesh well with highly factual, published, news articles. When it comes to news articles, an excerpt that is shared will very often be the most “important” aspect of the work – but that importance will derive from the uncopyrightable factual content, not the expression. It is not the “heart of the work,” but a piece of the factual skeleton upon which the expression hangs.

Third, the court discounted the value of robots.txt files to provide permission.  Robots.txt files give sites a tool to disallow (or allow) certain crawlers, and other courts have sensibly looked to robots.txt to find whether there was permission to include a site in search results. The default is permission, which a file can remove. This voluntary system has been vital to the development of tools to help users find material online. While the court does not directly disagree with the leading robots.txt cases, it finds they do not apply here.

There’s a lot more in this 91-page ruling, much of it troubling. Meltwater has said "We're considering all of our options, but we look forward to having this decision reviewed by the Court of Appeals."

Related Issues