For those in Northern Ireland whose only sources of local political goings-on are what elected representatives share on social media platforms, one could be forgiven for thinking that the political class in this country are in the midst of an intellectual renaissance. Councillors and Members of the Legislative Assembly (MLAs) previously known for their concise, or at times barely coherent, communications styles have begun posting detailed, erudite, and seemingly well-argued statements with impeccable grammar and syntax on social media. Unfortunately for those still with faith in politics, the posts you are arguing under are likely not written by your elected representative, but rather copied and pasted from an AI app based on a short, barely-thought-out prompt.
Generative artificial intelligence, specifically platforms like ChatGPT or Grok (the AI chatbot built into X), is increasingly becoming the ghostwriter behind many public representatives’ online personas. While AI tools are certainly powerful aids for summarising lengthy documents and extrapolating information from dense sources, their growing use in public-facing communications poses concerns to the democratic ideals that much of the population still espouse.
Politicians who present themselves online as appearing vastly more knowledgeable and articulate than they actually are in real life are misleading voters, cultivating a self-image that is entirely at odds with their actual capabilities. The average voter should be entitled to trust that the individual they elect genuinely possesses the traits implied by their social media posts, whether of breadth and depth of knowledge, moral clarity, or any other characteristic that is easy to fake to an extent with the help of AI.
This is why, on platforms such as X and Facebook, and on other political communications, you might see some small print stating that said communication is “promoted by Unknown Party Official on behalf of the Chameleon Party” or similar; guidance from the Electoral Commission and the Advertising Standards Authority states that when an individual’s public facing communications are being managed and created by someone else, this should be disclosed to the public. Such norms have not yet become realised when it comes to AI, still viewed as a tool rather than an author in its own regard.
When it becomes evident, as it has done in several instances to those of us tuned in to such things, that certain individuals cannot replicate in real life the clarity of thought and cogency of communication which they deliver online, voter confidence becomes eroded and members of the public who pick up on this simply feel lied to.
In Northern Ireland, where confidence in political institutions has historically been very low, it is disappointing to see that such unfortunate trends are continuing because of, rather than in spite of, the many new tools that exist to make our lives easier. The public are already sceptical about the authenticity of public figures, particularly politicians, and especially Keir Starmer, whom many believe to have opinions so fluid that they would be useful during a hosepipe ban. Using ChatGPT to parrot talking points on social media rather than actually engage with conversation meaningfully does little to assuage the belief that politicians are little more than puppets.
As numerous recent social media posts show, reliance on AI-generated content results in repetitive, hollow communications that do not introduce fresh ideas. By design, these AI chatbots blend material from extant sources and regurgitate information in an order they predict will most closely match what they expect you are looking for, based on the inputed prompt. The next time you see a Facebook post from one of your local councillors that is several paragraphs long, think for a minute if they are saying anything with substance at all, or if it is all information you have heard before elsewhere. I guarantee the latter is true.
Such habitual use of AI by individuals who are likely more than capable of coming up with their own viewpoints and putting time and effort into crafting a meaningful statement does little other than encourage intellectual complacency among representatives. It is clear, though, why they do it. It is very easy, and the majority of the electorate simply cannot tell the difference. One former party officer in Northern Ireland who I spoke to on this topic commented that it is simply quicker than having an original thought.
Worldwide, ChatGPT usage aligns with the age demographics that you would expect: over 50 per cent of regular users of the app fall within the 18-34 age group, with those over the age of 65 accounting for only around 5 per cent of users. There are clear tell-tale signs that something has been written with AI, and those who come across them often recognise them instantly. Social media in general is full of AI-generated posts, so much so that some people believe the internet is dead, devoid of real life and full of bots talking to each other.
For older individuals who are unable to keep up with the pace of change of technology online — and I don’t blame you, for the pace is increasing exponentially — it just wouldn’t cross one’s mind to think that the name of the person on the screen doesn’t match up with the content they are putting out. But the reality is that an increasingly large portion of content online is AI-generated, and the fact that elected representatives are contributing to this is worrying.
It should also be worrying others within the political parties of prolific AI users. Interest in politics by young people in the UK is at historically low levels, especially so in Northern Ireland. Using AI to create political communications instead of attempting to meaningfully engage with disaffected groups is not likely to win them over, and any serious discussion within parties on winning the youth needs to talk about this, as it is a serious issue at local levels.
I urge you to stop taking the easy way out of public communications and actually put some thought and personality into your messaging
The reason the underdog mayoral candidate for New York City Zohran Mamdani was able to was able to secure the Democratic nomination recently was because he explicitly engaged younger voters by utilising tech the way they do; even though young people are a minority force when it comes to voting, they have an increasingly outsized influence due to technology and the internet, which can be effectively leveraged.
Young people are not immune to having the wool pulled over their eyes, of course — however, they know when things online are not what they may seem, and will laugh at individuals making clear and obvious uses of AI. If you, reading this, are a public figure seeking (re)election in the next few years, I urge you to stop taking the easy way out of public communications and actually put some thought and personality into your messaging. Don’t simply post AI-generated paragraphs on social media because you now get more Facebook likes from people who attend constituency meetings. The next generation are looking at you, wondering why you are being so lazy, and they can’t wait for someone real to replace you.
