Automation: the force modernity that drives forward progress towards a better future. Or is it? The risks of automation have been well documented by researchers in universities. The most terrifying of these researchers is the one done by two Oxford University researchers, Carl Frey and Michael Osborne who both suggest that the number of jobs at risk of automation in the United States of America is 47%. The fact that nearly half of all jobs are at some risk of automation should alarm policymakers. However, the research is only for American employment. What would be the effect of automation and AI on British politics?
Research conducted by the Think Tank Future Advocacy has analysed the effects of automation by different British parliamentary constituencies to look at the risk of automation by each parliamentary constituency and their research identified that the constituency most at risk of automation was none other than John McDonald, the shadow chancellor’s seat of Hayes and Harlington. What I have done is combined their analysis of the various constituencies and the parliamentary vote for both the Conservatives and Labour in 2017, the UKIP vote in 2015 and the EU referendum results in 2016. The findings I believe are a good indication to look at the current state of politics in the UK.
Labour, the party of the workers and trade unions should have a strong correlation towards its vote being in areas at risk of automation. Yet, overall there was a negative correlation between Labour vote share and seats at risk of automation.
The graph which plots Labour vote by constituency versus constituency risk of automation shows that Labour wins more votes in parliamentary constituencies where there is a lower risk of automation. That means Labour won votes in areas in the 2017 election where jobs are less likely to be automated. If Labour wins in areas where jobs are less likely to be automated, then it must be the case that the Conservative party have an even stronger negative correlation between Conservative votes and constituency automation risk, right? Wrong.
It turns out that the Conservatives have a strong positive correlation between Conservative vote by constituency and automation risk. This means the Tories won a larger number of votes in the 2017 election in areas with a greater risk of automation.
So, what is going in British elections for the Conservative voters to be more threatened by automation that the Labour Party? This research highlights underlying trends within British politics, that demographic voting trends have changed in Britain over the last two decades.
Firstly, the class is no longer a good predictor how people will vote in an election. Labour, the party of trade unions and the working class only received 44% of the working class vote in 2017 compared with 39% of the middle-class vote. In contrast to this in the 1997 election, Labour won 57% of the working-class vote and just 34% of the middle-class vote. Labour led by Jeremy Corbyn won more middle-class voters than Labour led by Tony Blair. The underlying trend here is that Labour over time has become more cosmopolitan, urbanised and diverse in its kind of voters, getting voters from young metropolitan graduates alongside ethnic minorities, public sector workers and women, whilst the Tories have concentrated greater support from older voters generally in smaller towns. The largest single predictor of how someone would vote in the 2017 election was not class but age; 63.6% of under-thirties voted for the Labour Party. Compare this to the 63.5% of over the sixties voting for the Conservative Party and it becomes clear that class is now a weak predictor of how people vote.
Secondly, it’s not entirely the case that Labour represented constituencies are overwhelmingly less likely to be threatened by automation. My research found a sort of paradox between Labour constituencies that are less threatened by automation and constituencies with a greater risk of automation. I have called this “The Eagle Paradox”. Why the Eagle Paradox? Because the Labour MPs of Maria Eagle and Angela Eagle represent constituencies on either end of the automation spectrum. Maria Eagle’s seat of Garston and Halewood is ranked 6 on the list of constituencies with the most risk of automation, whilst Angela Eagle’s constituency of Wallasey stands at 562 out of 632 seats. This illustrates the more complex nature of Labour’s constituencies; on the one hand, Maria Eagle represents the more traditional Labour constituency with large amounts of manufacturing typical of Labour heartland seats in the past. Angela Eagle represents the sort of constituency that Labour has improved in over the last two decades in seats which are a more middle class, cosmopolitan, urbanised constituencies.
Thirdly, a more disturbing tendency found in the research I conducted was that there was a strong link between Right-wing populism and constituencies. In the 2015 election, the UKIP party did significantly better in constituencies where jobs are at greater risk of automation.
The UKIP in 2015 was correlated with constituencies with a greater risk of automation. The Conservatives at the 2017 election targeted UKIP voters who voted for Leave in the referendum s part of their electoral strategy. As a result, this would be a good indication of the reasons why Conservative vote is correlated with stronger support in areas with a greater propensity for automation because the Conservatives won UKIP voters. The nationalism and Right-wing populism behind the Conservative coalition of voters explain in part the reasons why the Conservatives are doing better in areas with a greater risk of automation. The Conservatives in part have successfully rebranded themselves as a populist, pro-Brexit, nationalist party whose voters are socially conservative, older and traditional in their outlook. The strongest positive correlation between parliamentary constituencies does not come from political parties but from the European Union referendum result; constituencies, where the Leave vote was stronger, had a very strong correlation with parliamentary constituencies at risk of automation.
There is, therefore, a potential clear link between Right-wing populist movements on the one hand and automation; that area where jobs are at a greater threat of automation are vulnerable to Right-wing populist. This connection between Right-wing populism is not just limited to the UK. The New York Times piece, Robots Can’t Vote, but They Helped Elect Trump, Thomas Edsall cites research conducted at MIT that automation helped push people to the populist-Right of the political spectrum. This he argued was a significant factor to the reasons why Donald Trump was elected in 2016 election. This gives us a warning that automation pushes voters to the clutches of the populist-Right.
 Carl Benedikt Frey, Michael A. Osborne, THE FUTURE OF EMPLOYMENT: HOW SUSCEPTIBLE ARE JOBS TO COMPUTERISATION?, Oxford Martin School Publications, https://www.oxfordmartin.ox.ac.uk/downloads/academic/The_Future_of_Employment.pdf , September 17, 2013
 Matthew Fenech, Cath Elliston, and Olly Buston, The Impact of AI in UK Constituencies: Where will automation hit hardest?, Future Advocacy, http://futureadvocacy.com/publications/, 2017
 Vote share figures from the election were accessed in Vyara Apostolova, Lukas Audickas, Carl Baker, Alex Bate,Richard Cracknell, Noel Dempsey, Oliver Hawkins, Rod McInnes, Tom Rutherford, Elise Uberoi, General Election 2017: results and analysis, House of Commons Library, BRIEFING PAPER Number CBP 7979
 Ibid, page 43
 Ipsos-Mori, How Britain voted 1997, 31st May 1997, https://www.ipsos.com/ipsos-mori/en-uk/how-britain-voted-1997
 Chris Curtis, How Britain voted 2017, Yougov, https://yougov.co.uk/news/2017/06/13/how-britain-voted-2017-general-election/
 UKIP vote accessed at, UK parliament.net; Olivia Hawkins, Richard Keen, Nambassa Nakatude, House of Commons Library, General Election: Results and Analysis, Briefing Paper, Number CBP7186, 28 July 2015. The outlier in the graph is the constituency of Clacton that UKIP managed to win in the 2015 election.
 Leave Vote in EU referendum accessed Map Logic.com (https://www.map-logic.co.uk/products/eu-referendum-results-by-constituency) ; Martin Baxter, Electoral Calculas.com.
 Thomas B. Edsall, Robots Can’t Vote, but They Helped Elect Trump, New York Times, January 11th 2018, https://www.nytimes.com/2018/01/11/opinion/trump-robots-electoral-college.html
All graphs were created by the author himself with help from research