Is There a Privacy Risk in Google Flu Trends?
The New York Time – 13/11/2008 By Miguel Helft
When Google released its Flu Trends service earlier this week, the Drudge Report flashed a headline that read: “SICK SURVEILLANCE: GOOGLE REPORTS FLU SEARCHES, LOCATIONS TO
Google sought to avoid this kind of reaction by talking about how Google Flu Trends protects the privacy of its users. The service relies “on anonymized, aggregated counts of how often
certain search queries occur each week,” Google said.
Still, the worries persist. On Wednesday, two advocacy groups, the Electronic Privacy Information Center and Patient Privacy Rights, sent a letter to Eric Schmidt, Google’s chief
executive, raising privacy concerns: “The question is how to ensure that Google Flu Trends and similar techniques will only produce aggregate data and will not open the door to user-specific investigations, which could be compelled, even over Google’s objection, by court order or Presidential authority.” The letter went on to challenge Google to publish the techniques it has adopted to protect the privacy of search queries used for Flu Trends.
There is no doubt that there are longstanding and legitimate privacy concerns about the collection and storage of search data by companies like Google, Yahoo and Microsoft. They all retain logs of the searches conducted by millions of people for varying periods of time. The logs include the search terms used in a query, the I.P. address of the computer that sent the query and a cookie associated with that computer. Those search logs could be misused by the companies, and they certainly can be subpoenaed by the government or sought by private litigants in civil lawsuits. More than two years ago, The New York Times showed how the data could be used to identify the individuals behind certain queries, at least before the data is partially “anonymized.”
But it is hard to see how Google Flu Trends changes any of that it any meaningful way. To create the service Google did not collect any new data. It did not change its policies on how it handles search data. It did not change the timing on when it anonymizes the data, a process that makes it harder, though perhaps not impossible, to associate multiple queries with the same user or computer. That still happens at some point between nine months after a search, when Google scrambles I.P. addresses, and 18 months, when it deletes the cookie associated with a query. And Google did not expose the raw data about flu-related queries to the outside world — it simply allowed anyone to chart their rise and fall in every state. In analyzing aggregate data that Google already has, Flu Trends is no different than Google Trends or Insights for Search, two related services that let anyone track the ebb and flow of any search term or phrase. You, or the government for that matter, can already see which states are generating more searches for SARS, Viagra or the flu. You can track the relative frequency of searches for “bomb making,” which appear to be particularly popular in Pakistan,
though New Zealand is a close second. (In the United States, “bomb making” queries were a
favorite in Ohio, Washington and Michigan, at least as of Thursday evening. I wonder what the government makes of that.)
The point here is that Google Flu Trends highlights an issue that has existed since the day search engines began collecting and storing queries. Search data is very powerful, very private and potentially very revealing about a searcher’s intentions. And without better legal safeguards, it can certainly become the target of government investigations and other intrusions, potentially putting people’s privacy at risk.
I talked to Marc Rotenberg, executive director of the Electronic Privacy Information Center, and he agreed with me to a large extent. “I think Google Flu Trends makes clear an ongoing
problem with the use of Google search which has not been resolved,” said Mr. Rotenberg, one of two co-signers of the letter to Mr. Schmidt.
But he insisted that, by highlighting the unresolved problem, Google Flu Trends made things worse. “I think this does change the game,” he said. “Google is inviting public health officials to make more of search information. In doing so, it has heightened the privacy risk to people who search for health information online.”
Or perhaps it has just let people know that if they are concerned about the privacy of those searches, they should use a computer at their local library or some other place where tracking them would be harder.