Google tracks flu (and you?)
There are many aspects of Google I don’t like (such as StreetView) but Google’s latest initiative is an early warning system for flu outbreaks. It’s called Google Flu Trends and is no doubt part of their efforts to break into the health industry, which I’ve blogged about before. Google is suggesting that its new tool could help to spot flu trends up to two weeks before being reported by the Center for Disease Control and Prevention in the US. At the moment, Flu Trends is US specific.
Apparently, Google have noticed a spike in search terms up to two weeks before officials announce a flu outbreak with users suffering from flu-like symptoms searching terms like “flu symptoms”. This is when Google’s “I know everything you search” capability is put to good use! So Google’s Flu Trends analyses the search terms in real time and creates graphs and maps showing clusters of flu outbreaks.
Normally of course, people trot off to the doctor or hospital and it can take time for data about flu trends to be accumulated, analysed and officially reported. Google is clearly leveraging collective intelligence with this new tool and using search traffic for flu surveillance.
From a privacy perspective, individual searches are not identified according to the tool’s Protecting User Privacy statement. But being the suspicious type, I do wonder what the relationship is between Google Flu Trends and Google Health. I’ve blogged before about what I see are the potential issues around Google Health here and here.
Of course, the PR spin on all this is that Google Health and Flu Trends put power in our hands. We can be in control of our own health records and be better informed – imagine that you login to Google Health and it alerts you to a potential flu epidemic in your area. You are forewarned.
But you know: something nags at me. Google can in fact whip up a list of people who search for particular terms (identified by IP address and/or cookie value). Google has admitted this. So I’m not sure their privacy statement about not being able to identify people searching for keywords relating to flu holds. Just imagine for a moment that a person searches for “symptoms of AIDS” or “alcoholism signs”. Search terms and history reveal a lot of personal information and if linked to a particular user, there could be adverse consequences for employment, insurance or travel.
The Electronic Privacy Information Center (EPIC) shares similar concerns. EPIC sent a letter to Google’s CEO urging Google to be transparent and reveal the algorithm on which Flu Trends data is based and disclose exactly how Google protects privacy. The Drudge Report certainly sees a Big Brother aspect to all this when it released its reaction to Google’s new tool with the headline “Sick Surveillance: Google Reports Flu Searches, Locations to Feds”.
Whilst Google says it will anonymise search data, Google might want to read a paper called Trail Identification: Learning Who You Are From Where You Have Been, which states in part:
“Consider online consumers, who have the IP addresses of their computers logged at each website visited. Many falsely believe they cannot be identified. The term “re-identification” refers to correctly relating seemingly anonymous data to explicitly identifying information (such as the name or address) of the person who is the subject of those data. Re-identification has historically been associated with data released from a single data holder.”
My wild imaginings lead me to ponder – could Google Flu Trends, coupled with Google Health, lead to the following sorts of scenarios to name just two:
- insurance companies identifying areas as prone to certain diseases and denying health insurance
- employers using health data to screen candidates
Of course, you can give Google the flick and use an anonymising service like Tor, which defends against network surveillance that threatens personal privacy. It bounces your searches and communications around a distributed network. Check it out. I of course use it!