Google Inc., the world’s largest Internet-search provider, is seeking to black out portions of a transcript from a public court hearing that includes information on how it mines data from personal emails.
Google, fighting a lawsuit claiming its interception of emails amounts to illegal wiretapping, asked U.S. District Judge Lucy H. Koh in a filing to redact “confidential” information from the transcript, without being more specific. The main revelation at the Feb. 27 hearing was the existence of “Content Onebox,” used by Google to intercept emails for targeted advertising and to build user profiles, Sean Rommel, a lawyer for plaintiffs, told the judge at the time.
Google’s latest move to keep records in the case out of public view comes as Koh is weighing a request by companies including National Public Radio, New York Times Co. and Washington Post Co. to unseal other key documents filed earlier that the company contended were too sensitive to be revealed.
According to Rommel, Google in September or October 2010 moved Content Onebox, which previously scanned stored emails, to “the delivery pipeline,” to extract data before users received the messages. Live or unopened communications are more protected under the federal Wiretap Act than stored ones.
Rommel told Koh that Google made the change after determining that Content Onebox couldn’t extract information from emails that hadn’t been opened or deleted, or that were opened on an iPhone or sent through an email provider other than Google.
After the switch to scanning unopened emails, Google began creating profiles of people “from which they could extrapolate additional advertisements,” Rommel said, without giving details of how Content Onebox works.
“Google itself has admitted and declared that the location and the timing of Content Onebox’s existence is proprietary, it’s secret, it’s unknown,” Rommel said at the Feb. 27 hearing. “There’s not a single disclosure in the record which identifies that there’s a content extraction feature occurring in the delivery process, which would be the interception.”
Google’s attempt to black out portions of the transcript that haven’t been publicly reported in press accounts is at odds with statements from Michael Rhodes, the company’s lawyer.
“We came here today and we unburdened the court of any sealing effort,” Rhodes told Koh at the Feb. 27 hearing. “We agreed that all of the material that had been previously designated confidential could be aired in the public courtroom so that those folks back there in the media could see that Google has nothing to hide here.”
Google officials didn’t immediately respond to an email seeking comment.
At the hearing, Rommel discussed a predecessor to “Content Onebox” called “Cat II mixer,” which he said was used for Google’s advertising business. Rommel also claimed Google’s interception of emails violates opt-out agreements that educational institutions or small businesses signed to shield themselves from the company’s scans.
Google has said in court documents that its practice of scanning emails and use of scanned information isn’t a secret.
“However, the mechanics of how Google performs those processes are sensitive, both for user security reasons and for competitive reasons,” Google argued in a court filing. “The specific techniques that Google uses to implement Gmail’s processes, the systems and infrastructure it uses to apply those processes, and the sequences in which it applies those processes are the result of over nine years of development by Google’s engineers.”
Koh in September denied Google’s bid to dismiss the case. The judge rejected Google’s argument – which it continues to make – that Gmail users agreed to let their messages be scanned when they accepted subscription service terms and privacy policies.