Does GitHub seem a likely host for illegal porn downloads?
Takedown Piracy LLC, an agent that sends DMCA takedown notices on behalf of various content publishers, sent Google a DMCA complaint back in December (please note that link contains some not-safe-for-work text). Most of the links targeted by the complaint appear to be illegitimate copies of material published by Wicked Pictures, an adult video company.
But among the URLs listed in the complaint are a slew of GitHub projects, none of which have any discernible relationship to Wicked Pictures or its content. Among the project repositories targeted for being yanked from Google are works from OpenSuse, Knockout.js, Facebook (the Rebound.js project), Netflix (the Lipstick project), and even a GitHub help page entitled "Dealing with non-fast-forward errors."
The most likely explanation for why GitHub projects would get flagged in this dragnet is some variety of automated DMCA generation script that's fed by keywords, then submits to Google with little or no human oversight. Many of the keywords flagged in the projects in question -- "satyr," "darling," "lilith," and so on -- match the names of some of Wicked Pictures's productions. One of the broadest matches, though, "cargo," also matched the name of the Rust language's package manager.
On the plus side, the short-term damage done by the takedown notice seems to be minimal, since the takedown affects only the specific URLs described in the notice. In other words, while the main page of a repository may be blocked from a Google search, all of its other subpages still show up -- and the project remains on GitHub, untouched. (Any searches conducted through Google that are blocked are listed in a ChillingEffects.org report link at the bottom of the search page.) That said, the owners of each individual project may still have to separately petition Google to have the deleted links restored.
Automated, shotgun-spread DMCA takedowns have long been a problem for both content owners and those affected by such takedown claims. Many of the problems revolve around lack of direct human supervision for automated takedowns. Earlier this year, Sony demanded that the open source film "Sintel" be taken down from YouTube, despite the fact that Sony has no copyright claims over the project. InfoWorld's Simon Phipps believed that in such a case, YouTube's Content ID algorithm was provided by Sony with something that produced a false positive that never had any human supervision to verify it. (The block has since been lifted.)
Likewise, back in 2003, the Recording Industry Association of America sent Penn State University a takedown notice claiming the university's FTP site was illegally distributing songs by the musician Usher. Apparently the takedown notice had been automatically triggered by the presence of an MP3 file ("an a capella song performed by astronomers about the Swift gamma ray satellite") an item bearing the name of a professor emeritus at the university: Peter Usher. The RIAA apologized, claiming "a temp employee made a mistake and did not follow RIAA's established protocol" for the takedown notification.
GitHub wasn't petitioned directly with a takedown notice in today's case, but DMCA complaints against individual GitHib repositories are filed regularly. GitHub keeps track of all such notices as well as counterclaims filed against them, though the company refrains from reports on whether any given claim was in the right.