In 2013 the whistleblower Edward Snowden revealed a surveillance programme called PRISM, within which the United States National Security Agency (NSA) had accessed and analysed the metadata from phone calls, emails, and other digital data stored by Verizon and nine internet companies. This article seeks to show that surprisingly deep insights can be gleaned from metadata by applying a range of easily available network-analysis algorithms to a body of metadata generated by another government. The source is the British State Papers (now digitized at State Papers Online), which contain 132,747 unique letters from the period between the accession of Henry VIII and the death of Elizabeth I. An analysis of this archive shows us that we can observe not only broad patterns of communication but also anomalous behaviour, and can make predictions about people likely to be trading in conspiracies or illicit intelligence. These discoveries demonstrate the power of such methods for the study of history. This power, however, is merely a shadow of that wielded by government bodies and private companies and therefore the findings also act as a warning about the potential uses and abuses of the metadata we generate with each of our digital communications.The article begins
‘We kill people based on metadata.’ General Michael Hayden, former director of the NSA, 2014.
On 5 June 2013 the Guardian newspaper published an exclusive article, based on information leaked to them by Edward Snowden, that revealed a large-scale effort by the United States National Security Agency to collect domestic email and telephone metadata from the US telecommunications company Verizon. In the following days it would emerge that this was part of a more widespread and systematic programme, named Prism, which allowed the Agency to gain access from nine Internet companies to a wide range of digital information on foreign targets operating outside the United States. The NSA defended itself against allegations that its activities contravened the Fourth Amendment by arguing the difference between data and metadata: that the contents of the phone calls, emails and other communications remained private. Citing a Supreme Court ruling from 1979, however, they maintained that Americans had no reasonable expectation that the metadata produced by their telephone and Internet communications – the names of those making and receiving such communications, the times and dates when they were made, and the geo-location of each party – should remain private. Now, in the era of the Trump administration, there is renewed unease. Trump’s appointment to head the CIA, Mike Pompeo, has previously stated that ‘Congress should pass a law re-establishing collection of all metadata, and combining it with publicly available financial and lifestyle information into a comprehensive, searchable database’.
To a layperson it is perhaps unclear why this should be of concern. How much can metadata really reveal? This essay seeks to show that surprisingly deep insights can be gleaned from metadata by applying a range of easily available network analysis algorithms to a body of metadata collected by another government. Our source is the State Papers held at the British National Archives, specifically the correspondence held in this archive dating from the accession of Henry VIII in 1509 until the death of Elizabeth I in 1603, which comprises 132,747 unique letters (duplicates have been excluded). Analysis of this archive shows that we can observe not only broad patterns of communication but also anomalous behaviour, and can make predictions about people likely to be trading in conspiracies or illicit intelligence. These discoveries demonstrate the power of such methods for the study of history. This power, however, is merely a shadow of that wielded by government bodies and private companies. The following pages also act as a warning, therefore, about the potential uses and abuses of the metadata we generate with each of our digital communications.