Google aims to bol­ster pri­vacy

Los Angeles Times - - BUSINESS BEAT - By Makeda Easter makeda.easter@la­times.com

Google an­nounced two new steps to pro­tect user pri­vacy — mov­ing to scrub per­sonal med­i­cal records from search re­sults and halt­ing its long-stand­ing pol­icy of scan­ning emails to de­liver tar­geted ads.

Pre­vi­ously, Google sur­veyed the con­tents of emails to pro­vide per­son­al­ized ads to users of its free Gmail service. Although pay­ing Gmail cus­tomers were never sub­ject to such scan­ning, Diane Greene, a se­nior vice pres­i­dent at Google, told Bloomberg that there was con­fu­sion about the pol­icy among busi­nesses that pay for its service.

“G Suite’s Gmail is al­ready not used as in­put for ads per­son­al­iza­tion, and Google has de­cided to fol­low suit later this year in our free con­sumer Gmail service,” Greene said in a Friday blog post.

The shift comes as Google tweaked its search en­gine to help hide re­sults that in­clude “con­fi­den­tial, per­sonal med­i­cal records of pri­vate peo­ple.” The change was also first re­ported by Bloomberg.

The pol­icy ap­pears to be aimed at pre­vent­ing data posted by in­di­vid­u­als from ap­pear­ing in search re­sults more than data from in­sti­tu­tional health­care providers, where pa­tient pri­vacy is pro­tected un­der fed­eral law.

“This kind of pol­icy can make it harder for ma­li­cious leak of med­i­cal records,” said Peter Swire, a pri­vacy ex­pert and pro­fes­sor at Ge­or­gia In­sti­tute of Tech­nol­ogy.

Google has pre­vi­ously taken steps to mask search re­sults that in­cluded in­di­vid­u­als’ fi­nan­cial in­for­ma­tion and ex­plicit pho­tos up­loaded with­out a per­son’s con­sent.

Search en­gines such as Google care­fully weigh which types of in­for­ma­tion they ex­clude, said John Verdi, vice pres­i­dent of pol­icy at the Fu­ture of Pri­vacy Fo­rum. Mas­sive breaches in health­care data over the last few years — com­pro­mis­ing more than 100 mil­lion records in 2015 alone, ac­cord­ing to fed­eral data — present a glar­ing pri­vacy prob­lem. But that could be more chal­leng­ing to ad­dress.

”So­cial Se­cu­rity data is straight­for­ward. It's the sort of thing that a com­puter pro­gram can read­ily iden­tify,” Verdi said. “Health records are not as eas­ily char­ac­ter­ized — it is a more com­plex task for a search en­gine that is go­ing to try to pro­gram­mat­i­cally iden­tify them.”

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.