Remember the “Google Forget” feature? Here is how it works as explained by Google

Google’s ability to forget is not the same as ours. It requires a calculated process rather than an incidental happening.

Yesterday, Google provided insight to a European government data protection working party on the handling of link removal requests from search results. This practice was initiated in June following an upholding in May of a Spanish man’s right to be forgotten (refer to our coverage).

Previously, the working group had given Google a questionnaire about the practicalities of such removals and had meetings not only with Google but also two other unmentioned American search engines. Following this, Google disclosed that it was dealing with these removals on a case-to-case basis, relying on the judgment of their recently-recruited workforce. Companies providing services to help people with link removals have started experiencing rejections, as reported by The New York Times.

In their cover letter, Google said that their approach would remain flexible and they anticipate continuing communication with data protection authorities. The letter outlined the criteria followed by their staff in granting or rejecting requests for link removal, which are similar to those set by the May court ruling.

Google highlighted several challenges they’ve faced in carrying out these removals, like the varied policies of EU countries regarding the full name publication in court documents. Google also finds it hard to determine comparable reputable news organizations online.

Google’s policy of notifying users when search results containing names have been modified due to legal actions, was also explained:

“We aim to notify users when there’s a possibility that their search results have been impacted by a removal, without revealing specific details about actual queries that were affected.”

The key takeaway was the company’s admission that it’s yet to devise a satisfactory method of translating the court’s order into a systematic calculation for identifying what belongs to public versus private interest.

“The decision-making process with respect to these removals is not yet automated. Every request demands individual scrutiny and this is done by humans. We have allocated a substantial workforce to this task, necessitating a sustained hiring effort to ensure enough resources are available for processing requests.”

To understand the scale of this project, consider the numbers: Google received about 91,000 requests within the first seven weeks of making the related form available and is still clearing the backlog. At present, it has approved 53 percent of requests, rejected 32 percent with justification, and requested additional information in 15 percent of the cases. Google has already reversed some decisions, including those impacting high-profile cases likes The Guardian. The House of Lords subcommittee in the UK recently described the court’s criteria as “vague, ambiguous and unhelpful,” reports the BBC.

source: Lucas Laursen/IEEE Spectrum