Will politics become a game of AI?

Cameron Conn

20 January 2022

Politics in recent years has been far more unpredictable and unstable than it was in the preceding period. Here are a few examples I’d like to focus on:

  • Russia’s annexation of Crimea. This resulted in the removal of   Russia from the GB and the resumption of a pseudo war be­ tween The West and Russia.
  • Donald Trump’s presidential success. The messaging used by Trump to get elected was unlike any seen before, focusing on antagonising any perceived threat rather than policy.
  • The UK “Brexit” referendum. Both sides of the aisle often used dataless emotion to secure votes rather than proper argument. The UK has since had 3 Prime Ministers in 6 years.

These turbulent events happened in consecutive years (2014-2016), a statistically unlikely triad that was arguably the first of its kind since the Berlin wall demolishment to Soviet Union dissolution of 1989-1991.

A common view is that this political unrest was caused indirectly by the financial crash of 2008. The argument is that the crash created comprehensive discontent among the world’s population; driving people to desire radical alternatives and vote for the most extreme change no matter the cost. While I accept this  in part, I believe the dominant factor is interconnected with the ascension of two converse phenomena:

  • How we use internet data
  • How the internet uses our data

The introduction of the iPhone in 2007 started a revolution in how we access internet data. Let’s take the UK as an example. In 2007, only 70% of the public had a home computer of any kind[1] and very few people had a smartphone. By 2011, 87% had a touch screen smartphone with internet connectivity [2]. Suddenly, the world was consuming data like never before. We were clicking, reading and watching our way through the day, rendering the 20th century view that a good memory is needed to be intelligent totally retrograde. Today, someone who can browse the web effectively can retrieve information at a rate which would put even the best eidetics to shame.

However, this utopia has its flaws. We are now completely at the mercy of the data we consume. Our opinions, policy preferences and general political views are of course a consequence of the articles, news broadcasts and social media posts we’ve seen. This unfortunately means that incorrect data or appeasing data (“clickbait”) can shape a population’s view beyond repair, resulting in political instability. The world was ready to be exploited.

Fortunately for the pessimist, humanity always finds a way to exploit where exploitation is possible. Enter, stage right of course, Steve Bannon, Lee Cain and Dominic Cummings. There are other examples, but in my opinion these 3 are the Svengalis of the data-driven political age.

Bannon was the CEO of Trump’s successful presidential campaign. He famously used Cambridge Analytica to measure the public mood from Facebook-likes and smartphone data, before targeting calculated swing voters with calculated swing messaging. Interestingly, Cambridge Analytica’s methods are “to a large de­ gree based on the academic work of Michal Kosinski”, released in 2008, just after the iPhone.

Cummings and Cain should prob­ ably be grouped together. They were the architects of the successful “Leave” campaign in the Brexit referendum and Boris Johnson’s rise to power. Cummings is attributed with the misleading claim that the UK was sending “the EU £350 million a week”. The data point delicately ignored any money sent back from the EU to the UK, but modelling showed that the public would respond positively to the claim, even if they knew it was misleading. The assumed reasoning is that to dispute the claim, you have to explain that the transfer of money between the UK and the EU is bidirectional, making them seem tightly linked, and evoking a nationalistic desire to break that link. This is only hypothesis though, and it is quite remarkable that a data­ driven approach can create political messaging with a known outcome, without understanding how it creates that outcome.

Cain is the messaging king. He proudly boasts on the website for his new company, Charlesbye, to have developed the famous slogans “Take Back Control”, “Back Boris” and “Get Brexit Done”. He uses similar meth­ ods to Cummings to develop his messaging, even advertising them as “data-driven” and explaining they are “memorable, powerful and authen­ tic”. Both Cummings and Cain regularly employed  3rd  party companies to collect personal data and conduct surveys so they could develop their messaging. The majority of the data came from browser cookies, a method of tracking how people use the internet, what they look at, and how they respond.

As you can see, the work of our Svengalis not only exploits how we use internet data, but also my other salient point: how the internet uses our data (such as our Facebook-likes and cookies). The combination of the two has enabled background individuals to hold immense power over us, changing the course of major nations irreversibly.

My conclusion follows; I believe the dance between our data and the internet has directly lead to the political instability of recent years.

Who is to blame? It’s hard to say. The world has certainly taken to blaming the data fiends themselves, with Dominic Cummings and Steve Bannon in particular becoming villains in the eyes of many. However, is this just the media fighting back against what it views as “cheating”? After all, much of the media is de­ signed to influence, and techniques

honed over centuries seem to have been usurped by a few AI models and lots of data. Do we blame the devices we use? Certainly, without them much of this would be impossible, but at what cost? We could also turn against the software, such as cookies, or indeed software providers and data managers like Facebook. If the reader fancies a laugh I recommend watching “Zuckerberg explains the internet to Congress” on YouTube to see why it might be a while before politicians have the understanding to mount any reasonable attack.

Personally, I don’t think anyone in particular is to blame. The reality is that as time has passed, innovation has scaled exponentially and as a result regulations and society have struggled to keep up with the changing times. Some good changes are happening, such as Apple recently disabling IDFA (their cross­ application tracking technology) by default, but the outlook is definitely concerning. It may take major intervention to stop politics becoming a game of AI rather than a game of policy.



[1] Thomas Alsop. Uk households: Ownership of home computers 1985-2018 , Dec 2020.

 [2] S O’dea. Uk: Mobile phone ownership 1996-2018, Aug 2020.