There is a trend in non-fiction — in fact my editor has been on to me about this lately — to reveal things. Apparently, readers like to feel they’ve got the inside track, even when there are no secrets to uncover. Perhaps this drove Yasha Levine to call his new book Surveillance Valley: The Secret Military History of the Internet. It promises to shine a light on the close and ongoing relationship between state surveillance and Silicon Valley. There are two problems with this. First, most of it is not secret. Second, I don’t think it’s right.
For all its later ubiquity, the internet started life as a niche project funded by the US Advanced Research Project Agency (ARPA) in the 1960s to connect the tiny number of computers engaged in the computer science projects it supported. While ARPA was indeed part of the Department of Defence, much of its research, including this project, was unclassified. Indeed, the network was literally called the ARPANET. But it went far beyond this, argues Levine, and was always intended to be a large-scale surveillance tool, with monitoring of populations ‘built in’. Most historians of the early internet, reckons Levine, have glossed over this bit.
Fortunately, I know one of the people who helped develop the ARPANET, Professor Peter Kirstein. For more than a decade, starting in 1973, Kirstein ran the UK’s connection to ARPANET out of his office at University College, London (where he remains a professor of computer communications). So I phoned him, and he told me:
Unless everyone’s been lying to me for 40 years, the ARPANET was not intended as, or used as, a surveillance system. The 1966 concept could never include more than 64 locations, with four computers in each location. It took the invention of the Internet Protocols a decade later to envisage a capacity of millions of locations and billions of devices.
According to Kirstein — and this tallies with many other written accounts — ARPA’s main aim was to ensure that the US stayed ahead of the Soviet Union in science and technology. The hope was that connecting computers together would help academics work more effectively. Yes, that would include developing applications that would be of interest to the military — but the network itself was a means to that end. (In fact, when Kirstein hooked up to the ARPANET, the UK government wasn’t interested — only with great difficulty did he obtain small amounts of initial funding for the project from those sinister arms of the state: the Post Office and the National Physical Laboratory.).
True, the military sometimes used these connected computers to analyse data — but that’s about computation, which is not the same as surveillance. The ARPANET was a broadly open communications system, meaning it was also used by home computer clubs, radical libertarian meet-ups, academic researchers, and who knows who else — but it wasn’t being monitored. ‘In the UK, only people approved by a committee I chaired were permitted to use the network. Once a group was approved, I didn’t monitor what they put on the ARPANET,’said Kirstein.
There’s a softly conspiratorial view of the world which sees anything funded by government as corrupt or menacing. But large moonshoot research projects such as ARPANET are often funded by governments because they’re the only ones willing or able to take the risk. Here’s a secret: lots of researchers take government money and fill in all the forms, while making it up as they go along. Throughout, Levine has too much faith in the US government’s power of foresight, and too little in engineers such as Vint Cerf, who were able to access government research grants to conduct exciting research without becoming hapless dupes in the mega-machine.
Public knowledge is again transformed into exclusive insight when Levine takes on the anonymous web-browser, Tor. As several studies (and the official Tor website) have explained, this was originally a US naval intelligence research project that also received funding from the US State Department. Levine claims no privacy activist has ever explained why the US government funded Tor. I’ll give it a go: it was part of a strategy to spread democracy and American influence around the world, and, secondly, if more people use Tor it’s easier for US intelligence to use it safely, too.
Well why is it then, asks Levine, that people have been caught using Tor, including the founder of the drug website Silk Road? Answer: because the browser is good but not perfect. (Ross Ulbricht was mostly unstuck by human error in fact.) Levine also reveals that Tor is not safe if you log on to Facebook with your real username and password. Anyone working for Tor will happily reveal this bombshell to you.
For all that, Surveillance Valley remains worth reading. The military history of the internet is not secret, but nor is it popularly known. Levine is right that both governments and big tech firms have a shared interest in surveillance — one reason privacy rules have generally been lax. He surfaces several stories that deserve more airtime. Did you know, for example, that initial funding and research into Google Earth was from the CIA via a project called Keyhole? Sergei Brin and Larry Page bought Keyhole in 2004, when Google went public and the pair suddenly had more money than pharaohs. A neat section on the rapid privatisation of the internet in the late 1980s is a reminder of how many technical decisions about the net are made quietly behind closed doors. Levine’s a good writer too: despite the heavy acronym use that inevitably accompanies a book about government research projects, it’s highly readable.
In the end, though, Surveillance Valley demonstrates something different to what’s intended. To get ambitious infrastructure projects off the ground requires a lot of high-risk but steady investment. The secret this book actually reveals is how critical government funding has been in the research, development and building of the internet. This is something that our current elected officials might wish to reflect on. So might the current residents of Silicon Valley, when filing their next tax returns.