Monthly Archives: October 2016

The Algorithmic Truth

graph-algorithm-big-dataThere has been much written in recent times about post-truth politics, and much associated naval gazing as commentators, analysts and politicians themselves have tried to understand how to deal with all this. Leave campaigners in the UK promised £350m a week for the NHS; Donald Trump still thinks he opposed the war in Iraq; and Vladimir Putin claimed no involvement with the war in Ukraine. Populism, reactionism, anti-intellectualism – call it what you will, it’s certainly got currency.

Speaking of currency, the pound has taken a pounding since the Brexit vote, and overnight trading on the 6th-7th October witnessed a flash crash that traders have struggled to explain. David Bloom at HSBC put it that the pound had become the de facto opposition in Britain. ‘Sterling,’ he said, ‘has become a structural and political currency.’ I’d go further than that. It’s the algorithms providing the opposition. The algorithms governing the trading desks even scrape news feeds to see potentially influential stories (one commentator suggested that comments from Francois Hollande on the Brexit negotiations may have triggered the ‘flash crash’). They sense and learn, and respond, constantly scoring and valuing political decisions and the smallest market moves.

The algorithms then form their own truth. This isn’t post-truth politics, it’s absolute truth. And that is potentially a whole lot worse.

%d bloggers like this: