Could AI have won Jeremy Corbyn the election? – LoopMe

Mobile Marketing

Stephen Upstone, CEO and co-founder of LoopMe, considers what role artificial intelligence could have played in the recent UK general election, after the technology's successful introduction to US politics back in 2012.

It’s safe to say that politicians resort to all kinds of tactics to win an election. Following the recent surprise result of the UK general election, I begin to wonder why neither of the parties, Labour in particular, seem to have exploited artificial intelligence to fuel their campaigning, or at least their campaign advertising.

Both parties have acknowledged the importance of online, with Labour’s huge success on Facebook seeming to have played a part in the final result. While the ads themselves are sure to have raised awareness, the data they can collate on their audiences could and should have been used to inform campaign strategy.

For an example of how the candidates could have benefited, look back to 2012, when Barack Obama became the first President to embrace AI in his campaign. Working with data scientist Rayid Ghani, Obama consolidated all data from social media, marketing, polls and surveys, as well as door-to-door conversations, and stored it in one database. They then used AI and machine learning to predict four outcomes for each individual voter. How likely were they to support Obama, show up at the polls, vote for Obama, and change their mind on the election based on a conversation about a specific issue?

Based on the models built by Ghani and his team, the campaign ran 66,000 simulations of the election every evening and used these results to direct their army of volunteers. It allowed them to focus on swing voters, and informed them which topics would be most likely to trigger the change in opinion. Since then, it is likely all politicians in the US have all used AI to some extent – even Donald Trump.

In contrast, UK politics seems content to use information from inaccurate polling, taken from a tiny pool of users. A final poll before the election surveyed just 1,532 people, 0.00003 per cent of the 46.8m who cast their vote – and unsurprisingly, predicted the wrong result. Our refusal to adopt modern technology is leading to incorrect predictions, and ineffectual campaigning.

During the last general election, Labour failed to engage voters on the topics that mattered, and it’s not a stretch to think that was because data from polls and focus groups was delivering inaccurate insights into what people really wanted. Both parties faced a similar dilemma this election – even if it was clear which policies appealed to particular supporters, did they appeal to the wider electorate?

By collating data on their audience, either by pooling existing voter information or by running mobile ad campaigns to create new data sets, AI can be used to analyse information and determine which issues will resonate with certain people. For example, do women in Northern cities, aged 18-34, who are not university educated, care more about unemployment than their counterparts in London, who might care more about affordable housing? These are questions AI can answer.

AI technology can also be particularly helpful for galvanising younger audiences – a key demographic for Corbyn’s Labour party – as they are more likely to engage on mobile devices and often do not feel involved in traditional politics.

These insights can be used to put volunteers on streets, armed with a much better idea of what arguments could convince young women, middle-aged men and pensioners to vote for their party over the opposition. Alternatively, the results can also be used to serve online ads, tailored to each individual voter in real-time, and delivered based on their unique statistical probability of being won over by the ad.

While Labour steadily gained ground on the Conservatives since the election was first called, the insights gleaned from AI modelling could have been the extra push that swung the balance in their favour.  Knowing how to speak to voters, and knowing which issues are the most likely to influence each and every individual and using that to drive the campaign forward, could be invaluable to Mr Corbyn and the future of the Labour party.

Stephen Upstone is CEO and co-founder of LoopMe