Raytheon predict a RIOT: your ‘predictive’ data

Last week, Ryan Gallagher in The Guardian wrote about RIOT, an experimental project by the defence contractor Raytheon. A Cyber-Tracking tool, RIOT (Rapid Information Overlay Technology) analyses photos and the location data that may be attached to the photos that social network users take and post to services like Instagram or Facebook.  Using a graphical browser a reviewer can see the activity, associations and relationships between a subject and other online users. It could  allow analysis of where a subject went and how often – and how they got there and who they went with. The fact that the word ‘Rapid’ is in the acronym is interesting. Rapid because it’s a zippy application to use? Or ‘rapid’ because the data can be analysed in almost real-time as new check-ins happen, facilitating real-time tracking and even real-time prediction about where a subject is going next, and with who.

The potential for intrusion and the impact on people’s privacy is significant. The truth is many of us are not always where we’re meant to be; when we take an hour off to go to the dentist we may meander back to the office via Century 21 or The Gap, we occasionally tweet exes whilst under the influence late on a Friday night; deeper analysis may show that the duration it took you to get from home to the office may indicate your average speed was in excess of legal limits. Issue the speeding fine now!

“Data can always be interpreted in a variety of ways, and interpretation is in the eye of the beholder. If the beholder (of data) is looking at you this closely, the context is not a social one – they’re not nodding approval at your sushi reviews.”

Regular social activity from from Twitter, Facebook and Foursquare could be used to review a person’s activities, associations, movements and perhaps predict future movements and associations. Interesting use of publicly available data. If this story revealed a marketeer doing the same thing, we would be less concerned,  perhaps we’d sigh, and think that we’d just have to deal with more location targeted marketing. But, because Raytheon is a large defence contractor, it’s not a huge leap to think this is the sort of that will appeal to governments – worldwide. It would be comforting to believe that such a wide-scale initiative would only be used to catch bad guys, because of course, you and I have nothing to fear, after all we’re members of the innocent majority.

We owe ourselves some reflection on this, are you happy having your personal data analysed to this extent? If the data were used like many web companies use data, cumulatively, there are obvious benefits. Traffic on Fridays for instance – opening extra lanes, extending traffic light sequences in hot-spots or suggesting retailers look at opening hours could all help relieve congestion in focus areas.

One of the issues is social media may not actually be reflective of real-world relationships and real-world intentions or trends. We know people who interact excessively on-line do not in fact want to interact with the same people in the physical world. We often pay more attention to those we don’t know well; they have less in common with us, are outside our close circles of commonality and are often the people who tell us the most interesting things. And of course, can you program an analysis engine to interpret satire or sarcasm? Humour and satire will be a trait that could land you in the mire – or worse. So, do you want your most personal data analysed to this extent?  You should have the choice.

 External links & references

  1. GCHQ to trawl Facebook & Twitter for intelligence: The Guardian
  2. Electronic Privacy Information Centre
  3. Effective Counterterrorism and the Limited Role of Predictive Data Mining
Andy O'Donoghue

Andy O'Donoghue talks about technology, some say, too much.