Wired Magazine – UK
The US department of Homeland Security has commissioned a one-year contract to investigate the efficacy of using social networks to identify instances of bioterrorism, pandemics and other health and security risks.
It is paying Accenture Federal Services $3 million (£1.8 million) to scan the networks’ for key words in real time to see if growing threats or health trends can be distinguished. So if an individual flags up a nasty cough in a Facebook update, for instance, the software will be looking to see if key medical terminology is repeated in connected groups or from other individuals posting from the same location.
“This is big data analytics,” said John Matchette, managing director for Accenture’s public safety department, who admits the technique is yet to be proven. “In theory, social media analytics would have shown timely indicators for multiple past biological and health-related events.” Mobile data mapping has been used in the past to track and predict population movements following natural disasters and algorithms can use data to track disease hotspots after the event. However, this latest experiment could provide real time information to help stem disease spread, develop early warning systems and help emergency services coordinate react in a timely fashion
According to a company statement from Accenture, the software will constantly scan blogs, as well as the usual outlets, but not all networks and channels have been decided upon. It’s no surprise that national security departments monitor social networks to look out for threats (Paul Chambers’ arrest after a tongue-in-cheek faux bomb tweet threat being a perfect example of when that monitoring goes very wrong), however Homeland Security is already being sued by civil liberties group Electronic Privacy Information Centre and is under pressure to answer questions about setting up fake social networking accounts to search for key words such as “virus” and “trojan”. The department has been accused of violating the public’s free speech and constitutional protections against unreasonable searches. No one would disagree there needs to be better systems in place to monitor and protect against the spread of infectious disease, however how data is monitored to do this has come under fire.
“The information won’t be tracked back to individuals who posted it,” stated Matchette.
Not everyone is convinced. “Even when data is in aggregate, we don’t have any clear policies around how data will be used and how it can be traced back, including if and when there are signs of an illness outbreak,” Deven McGraw, director of the health privacy project at the Centre for Democracy and Technology, told WebProNews. “I think it’s a legitimate question to ask [Homeland Security] what the guidelines are for using this data. I’d prefer they have a plan in advance for dealing with this, rather than waiting.”
A statement on guidelines from Homeland Security — which has begun aggregating data from the Centres for Disease Control and Prevention and collecting urban air samples as points of reference — is somewhat vague, but does admit there is room to home in on specific persons of interest. Information that is already “accessible on certain heavily trafficked social media sites” is analysed without gathering personal specifics on an individual, “with very narrow exceptions”.