Saturday, 3 August 2013

Have You Ever Polled A Human By Mistake, Mr. Cameron?

That line, is of course, a paraphrase of the famous "Have you ever retired a human by mistake?" question asked by Rachael to Deckard, played by Harrison Ford, in the iconic 1980s film Bladerunner, arguably the greatest science fiction film of all time.

The central twist of the film is that neither Rachel or Deckard are aware that they are both in fact machines, replicants, who thanks to implanted memories believe themselves to be human. Bladerunner asks important questions about the human condition: what is it to be human? Do we have souls? Is free will just an illusion?

Such questions have been asked by philosophers for centuries, of course, but we are going to hear a lot more of them in the years ahead thanks to the rise of Big Data and predictive analytics. Essentially, the increase in computing power and ability to store data now allows machines to crunch through vast amounts of data and apply complex statistical algorithms to it, well beyond the capacities of the human mind to manage. The machine then learns a model from the data which can be used to discern patterns within it and in some applications, predict human behaviour. This machine learning is itself a branch of artificial intelligence, the holy grail of which is to create a machine with equal or better intellectual capabilities than the human brain. Some data scientists even believe that in future we will be able to predict human thoughts.

Machine Learning is not some Sci Fi fantasy, it is here, and the fact that you are reading this article on the web means that it has almost certainly been used on you: machine learning algorithms will predict what books to recommend you on, or what ads to show you on Google, for example.

Now is is being applied to politics, with the news that the Conservatives have hired the data expert Jim Messina, the man credited with winning the 2012 election for Barack Obama.

The central idea is that by applying gigantic sets of polling and behavioural data to an election campaign, you can tailor your election message individually to voters in order to get them to respond to your message. I thoroughly recommend anyone to watch this rather sinister video of the overly smug Messina to see what is in store for us on this side of the pond.

Well, anyone watching that piece doesn't need machine learning in order to know the thoughts of the Political Class if they think this is the way to win elections: the underlying contempt they have for the rest of us could not be more stark. We are little more than dumb machines and units of consumption, to be sold a packaged message, spun, used, manipulated and discarded by the new uberclass as they see fit.

Although Messina's data driven campaigning is frighteningly sophisticated, it is in essence nothing new but a continuation of the same poll and focus group driven politics we have all come to despise over the last ten years. Dreary and reductive, such ultra-segmentation of a political message treats us all as atomised individuals with little regard for a greater whole beyond ourselves. In essence the tailor-made campaign is the absolute antithesis of the leadership, belief and vision that successful politics - in the sense of actually changing the country for the better rather than just winning office - requires. It is very noticeable for example, that previous leaders who were obsessed with data driven campaigning - Clinton and Blair for example - achieved very little in office. Obama looks like another failure in this regard. So does David Cameron.

But will it work in Britain in 2015? Against the kind of mediocre candidates the political class tend to produce, almost certainly so. However, against a true leader with courage, vision and integrity, a man or woman who treats the people with dignity and respect rather than contempt, the results are somewhat less predictable. Big data, machine learning and predictive analytics are tools, nothing more. Like all tools, they rely on a master craftsman to be used properly. Better and more sophisticated tools can help you refine your craft, but they can not in themselves supply the inspiration to create something of lasting value or beauty.

Like all political parties, UKIP will have to adapt and use the tools of the big data age, but we must never lose sight of the fact that politics is a human activity, and human beings need romance in the true meaning of that word. Far more than big data, what politics requires and always has done, is a romantic vision of a better tomorrow.