Skip to main content

A new look at the people erasing themselves from Google Search

A new look at the people erasing themselves from Google Search

Share this story

If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement.

With the "right to be forgotten" now firmly established in European law, hundreds of millions of people can now ask to be delisted by Google, effectively erasing themselves from Google Search. A new report from The Guardian digs into who has been using the new feature, using information accidentally revealed in the source code of Google's recent transparency report. The new data covers the 218,320 requests that were made between May 2014 and March 2015, roughly three-quarters of the total requests, slightly less than half of which (101,461) resulted in a successful delisting. The data has also been published on GitHub, and is open for deeper analysis.

Google breaks the requests into five categories: private personal information, child protection, political requests, public figures, or serious crimes. The vast majority of requests (more than 95 percent) were for personal information, ranging from requests to delist a personal address to delisting a person's name entirely. By contrast, fewer than 200 child protection requests were received during the period. In a statement to The Guardian, Google emphasized that the codes were part of an internal project to categorize requests, which was ultimately deemed not reliable enough for public release.

Still, the new data sheds new light on the larger legal project of the right to be forgotten. Google has strenuously objected to the right to be forgotten on free speech grounds and has dedicated significant legal resources to fighting it in court. Those efforts have largely been successful in the US, but European Union courts have treated the ability to remove information from Google Search as an important privacy measure, particularly in protecting children and the victims of crimes from unwanted publicity. The new data suggests that those use cases, while legally compelling, represent a vanishingly small portion of the overall requests.

7/14 1:05pm ET: Updated to include the data's Github page.