Columbia University researchers claim 28% of Google’s URL takedown requests are invalid

A research paper published by Columbia University’s American Assembly and Berkeley claims nearly a third of the more than two million URL takedown requests Google receives daily may not be valid.

Using data Google gives to Lumen database — an independent third-party research project created to analyze DMCA requests and other complaints around online content — the researchers reviewed more than 108 million takedown requests and found that 28.4 percent “… had characteristics that raised clear questions about their validity.”

The research findings showed that 4.2 percent of takedown requests targeted websites that failed to include infringing content specified in the DMCA notice, while other DMCA notices were aimed at sites that had been shut down more than a year ago. Other concerns around the questionable takedown requests included improperly formatted DMCA notices, subject matter that did not fit DMCA takedown guidelines and potential issues around fair use laws.

Reported by both The Washington Post and TorrentFreak, the authors of the research paper claim the high volume of “questionable” requests is the result of automated DMCA notices.

Joe Karaganis, a co-author of the research paper, told TorrentFreak, “With notice sending robots talking to notice receiving robots, the step of actually looking at the targeted content often drops out of the equation. The main contribution of our study is to go back in to look at the targeted content and make those human judgments.”

TorrentFreak pointed out that Google acts on 97.5 percent of DMCA notices, which means a large majority of the questionable takedown requests are implemented.

The full 160-page research paper can be downloaded here: Notice and Takedown in Everyday Practice.