Of the thousands of planes in the air each day, which ones might be involved in government surveillance? This is a question that concerned Peter Aldhous from Buzzfeed News in several award-winning investigations. In 2016, he discovered mysterious patterns in Flightradar24 data and identified about 200 federal aircraft routinely circling the skies above the US – and observing the public space. A year later, he extended his research and used a machine learning model to automatically identify previously undiscovered spy planes – removing the tedious work of looking at thousands of flight paths by narrowing down the problem to a few candidates worth doublechecking.
In this keynote, Peter Aldhous will talk about his investigations, the benefits and challegens of using machine learning in journalism and what’s special about working as a science reporter for Buzzfeed News.
