Who voted for the machines?

At some point in the next decade, probably between 2015 and 2020, an important milestone will be reached. Machine to machine communications on the internet will exceed those made by humans.

The phenomenon of machine to machine data flows has gone largely unnoticed by legislators and the public. But it is part of the internet’s evolution. If stage one was its adoption as a mass platform, then stage two is the harvesting and exploitation of knowledge that this embracing of the internet made possible.  And this data does not just come from us, nor is it harvested only by us.

Since 2000 researchers at the Massachusetts Institute of Technology and the UK telecoms giant BT have been working on the concept of the ‘internet of things’ – the incorporation of chips carrying unique internet addresses into everything conceivable from dogs to lamp-posts and from cars to food packaging. The potential is huge. Already our critical national infrastructure, telecoms, utilities, transport, are all controlled by machines which communicate with each other. In the near future sensors will be everywhere, via smart meters inside our homes they will be able to know exactly where we are, and what we are doing at any time by using a combination of mobile data and information on power usage. The sheer volume of internet traffic is no less daunting. ‘By 2017, an insatiable demand for bandwidth from across the globe, most crucially from machine-to-machine connections, will result in more than 19 billion global network connections carrying 1.4 zettabytes* of traffic,’ says a spokesman for global technology company Cisco.

But before we start to hang out the bunting now would be a good time – perhaps our last chance? – to consider how desirable all this is. Where is the democratic accountability in a world increasingly run by computers and data flows with no input from humans? The recent revelations about Prism – the mass surveillance programme in which the US National Security Agency was found to be monitoring the world’s internet traffic using artificial intelligence systems – highlight these concerns. It’s bad enough if humans are gathering and using such information without our knowledge, but machines…who voted for them to control our lives?

But even more worrying than machines collating information about us is the fact that the same machines now make decisions based on that data. Professor Kevin Warwick, author of the book ‘March of the Machines’ and head of cybernetics at Reading University, points out that, already, 30% of the stock market buying decisions in the City of London are made by computers purchasing shares based on incremental prices differences around the world. ‘The computer can make a decision on whether to buy a stock that can make a company go out of business and that can then make an industry in a country collapse,’ says Warwick.

The growth in machine to machine data inevitably makes us think about the famous Turing Test of Alan Turing, the father of modern computing. In 1950 Turing argued that when we cannot tell whether the voice on the other end of a phone is generated by a machine or a human due to its responses, then machines will have reached a key stage in their development – and will be able to think.

But machines that are talking to each other will not care about the Turing Test. Computers have no emotion or conscience and thus will interpret data strictly in terms of the task they have been told to perform. So, for example, machines will report information on what is universally acknowledged as a crime in exactly the same way that they will respond to the activities of a political activist – under the machines’ rules they are both crimes.

Data integrity becomes everything, as for the machines to make the best possible decisions they have to be supplied with the best possible data. The fear will be that the collection of poor data in one area will have a disproportionate impact on data in another area. So the level of data collection will have to be exhaustive, triggering more fears about our privacy as our lives are scrutinised in ever more detail.

To its credit, the IT industry is aware of the issues. Macario Namie, vice-president of Jaspar Wireless, which among other things connects vending machines that self-order their own stock, agrees that there are ‘some issues to be worked out around privacy and how much information we want to give to third parties’. He adds: ‘The technology is the easy part, there are broader issues on who pays – and how much do we want to allow.’

Indeed. But the question is, how and when do parliaments and the general public get a real say in those issues?

*) 1 zettabyte (ZB) is 1000 billion gigabytes

Peter Warren   Michael Streeter

Peter Warren and Michael Streeter
Journalists and authors of the 2005 book ‘Cyber Alert: How the World is under attack from a new form of Crime’ and  ‘All that Matters: Cyber crime and warfare’ published in 2013 by Hodder & Stoughton.