When we're talking about computational approaches, I'm going to be specific and talk about political bots for the moment. Political bots are automated social media accounts. They're also automated accounts that could exist on instant messaging apps or through other communication technologies.
Right now we know there was substantial use of those bots in the U.S. election. A report written by Sam Woolley, who is based at the University of Oxford, showed some concrete evidence. In the report that Fenwick McKelvey and I wrote about the state of bots in Canada, we also saw examples of political bots during the 2015 federal election. We've started doing some initial work during the election that's happening in Ontario today. Those results are not confirmed yet, and we still have some more analysis to do, but there definitely are examples of automation being used. Not all of that automation is necessarily for voter suppression tactics or for things that we necessarily would be uncomfortable with in the election.
An example is most media companies use automated approaches to send out tweets and Instagram posts and Facebook messages all at the same time rather than one individual typing on each of these different platforms. That's the form of computational political messaging that we're pretty okay with. It is very difficult to say exactly how much to measure in a quantitative way. Voter suppression exists because it is very hidden and hard to trace so I can't give you specific numbers.