John Brockman of Edge interviewed me in London in March. The video of the interview, and a transcript, are now available on the Edge website. Edge runs big interviews with several dozen scientists a year, with particular interest in people who do cross-disciplinary work. For me, the interaction of economics, psychology and engineering is one of the things that makes security so fascinating, as well as the creativity driven by adversarial behaviour.
The topics covered include the last thirty years of progress (of lack of it) in information security, from the early beginnings, through the crypto wars and crime moving online, to the economics of security. We talked about how cryptography can help less developed countries; about managing complexity in big projects; about how network effects lead firms to design insecure products; about whether big data can undermine democracy by empowering elites; and about how in a future world of intelligent things, security may become more about safety than anything else. Finally I talk about our current big project, the Cambridge Cybercrime Centre.
John runs a literary agency, and he’s worked on books by many of the scientists who feature on his site. This makes me wonder: on what topic should I write my next book?
A very interesting interview Ross. On the issue of the internet of things and in particular regarding your example of automobiles where there is a complex chain of corporate responsibility for the components of the system. Do we just accept the status quo, do we wait for something major and consequential to happen and then react, or is there something by way of regulation that could and should be enacted?
I am getting somewhat annoyed and a little worried that there is pretty much a universal acceptance by the media, politicians, academics, business and the public that automated vehicles will in a matter of little more than a decade be so prolific that governments will ban human drivers from the roads.
Although not directly related to the security engineering, surveillance and control issues, it is interesting how easily it has been accepted that there is ‘carnage’ on our roads and AVs will reduce fatalities by at least 97%. This is obviously nonsense, but it is a nonsense that will put considerable pressure on governments to get drivers off the roads because the motor manufacturers, machine intelligence researchers and producers adhere to Stalin’s maxim of ‘when there’s no person, there’s no problem’. If and when it is introduced (level 4 and 5 automation), all the problems of the technology will be blamed on unpredictable human drivers. If it were ever possible to outlaw drivers and completely automate all vehicles (level 5), it would of course be the mother of all surveillance and control technologies. It would also make us extremely vulnerable for the reasons Ross gives when outlining the responsibility of automotive industry for the components of the system. Obviously AV technology will greatly increase these problems, but as yet there has been little serious debate about the risks and we appear to be swept away on a wave of technological determinism.
The problems of AV technology are legion and it may not in the foreseeable future be able to operate on ‘open roads’ even if, and it is a big ‘if’, there is a demand for it. Nonetheless, in reply to Mike’s questions, something major and consequential is happening and we need to react now. AV technology may be just the latest manifestation of ‘AI hype’, but we should not, as Alan Turing observed, underestimate how effective AI hype is in silencing disagreement and criticism.
I do sympathise! The IT industry hype merchants tend to lose sight of common sense once they get going. In the case of remotely piloted ground vehicles, a useful antidote may be this article by John Naughton.
I did a few posts to the Naughton article. It was the most sensible and indeed ‘obvious’ observation on AVs I have seen in the Guardian which is usually an avid supporter of the technology. I was amused by one of the posts that claimed Mercedes were solving the last few metres delivery problem by incorporating drones into their automated vans. Apparently, the drones will fly the parcels the last few metres! As I said, the problems with AV technology are legion but it seems unstoppable.
Oh, sure; like AI in the 1980s or smart dust in 2000? like string theory in the 1990s or quantum cryptography now?
I have a Nexus 5X, which I bought because I thought it might get security updates a bit longer than a Samsung thingy. Now google says it will stop security patches in September 2018. Tell me, if I’m writing navigation code now for a Land Rover that will ship in 2020, who will be writing security patches for that code in 2040, and how will it be paid for?
It is pretty obvious from the blurb generated by motor manufacturers, IT corporations, IEEE, lobbying organisations like Autelligence, etc., that they are looking at AV technology as another way of locking in people to a dependency on software for which they will have to pay through the nose. Again, from my exchanges through posting, lectures and talks I have given on the topic, I am struck by how many people are content that they will not ‘own’ vehicles and will be totally dependent on software suppliers. Most of them are young and do have difficulty when it comes to ‘ownership’ issues, or, as they would put it, think I have difficulties.
Bye the bye, I have adopted the term ‘autonomous vehicles’ because Autelligence et al insist that we must call them ‘self-drive cars’ and get quite upset if you refuse to do so. Controlling language, as Turing understood, is essential for AI types and they can get very angry if you refuse to use their lexicon. AVs are not of course truly ‘autonomous’ but I am prepared to start with AV to meet them halfway.
Eventually something nasty will happen and the regulators will start doing their job. That’s always what it takes: a Ford Pinto, a bleeding baby on the front page of the Daily Mail, whatever…
Perhaps, but by the time ‘eventually’ comes round it may be too late and we will be left with the equivalents of highrise flats, poor quality modernist buildings, closed and sold off railway routes. On the bright side, I will probably be dead by then.