Artificial intelligence’s impact on elections and democracy could be very real
One would have to have been living under a rock for the last couple of years to not have noticed the immense growth and attention paid to artificial intelligence (AI). Coined the ‘AI Spring’, the recent surge in attention and development in AI technology shows no signs of slowing down anytime soon. Despite the benefits undoubtedly outweighing the negatives, many argue, rightly or wrongly, that AI does not come without potential threats or dangers.
With elections in the UK, the US and other European countries all happening in the next 12-15 months, questions are already being raised about artificial intelligence’s possible impact on the state of elections going forward. Will it help preserve the sanctity of democracy or could it usher in even greater uncertainty in the electoral procedures of our nation-states?
With elections in major G7 countries all quickly incoming, let alone further by-elections happening in the UK, it is worthwhile assessing just how AI could impact elections going forward, how it could exacerbate issues with misinformation, but also reach new audiences with greater speed and efficiency.
AI has undoubtedly enabled levels of productivity and efficiency like we have not seen before. This increased level of productivity naturally brings benefits to businesses, employees and in this case, political parties and politicians seeking to get their messages out faster and with greater cost efficiencies. It is very likely that parties and their candidates will seek to use AI to hone their approach in order to optimise their campaigns.
AI will be used to simulate and ‘game’ the impact of their messages on the electorate, as well as to understand how to get the most out of the large sums of money they spend on advertising and marketing.
Both the UK General Election and the US Presidential Election could see the first time that AI is utilised by campaigns with hyper-personalised content for appeals being made to voters for their support. This high level of personalisation, without the need for human resources behind it, could be a game-changer.
This ability to automate hyper-personalisation throughout a party’s political content could see a huge shift in the levels of engagement with the public, something that will be needed if politicians wish to improve the levels of voter turnout in the coming elections (previous voter turnout: UK: 68.8% / US: 66.9%).
However, whilst hyper-personalisation may provide a great benefit for the political parties, this access to a greater level of information may raise the fear that AI might not be used in the best interest of the voter.
Artificial intellect, a possible risk
Naturally, where there is an opportunity for new technology to do good, there will also be those who intend to use it for harmful and detrimental means. Hostile powers, organised crime, and domestic groups on either side of the political spectrum all loom as threats to democracy during the time of an election. All stand to gain from destroying the authority of an incoming political leader or invalidating an election in the eyes of voters, especially those whose allegiances tend to swing in elections.
It seems likely that generative AI may potentially be used to create incredibly persuasive deep fake videos, audio recordings and written content that has the potential to disseminate inaccurate information, fabricate false news stories, and manipulate public sentiment. In the most extreme cases, it could be used to create smears and falsehoods – falsely showing candidates endorsing policies that will polarise and enrage some voters.
As it stands, the technology that is being used to assess and flag AI-made content, both positive and negative, has not yet matured as much as we would have hoped to prevent this content from reaching its intended audiences. This means that election commissions and those charged with protecting the sanctity of the electoral process will have to focus on the impact that the technology could have on the democratic process and seek ways to combat it.
The future of AI-centric elections
It seems improbable that governments will not contemplate implementing AI-regulation especially when it comes to elections in the near future. However, with the AI ‘cat now being very much out of the bag’ in terms of the evolution of the technology, it seems unlikely that nation-states will be able to fully rule out and outlaw AI’s impact on elections going forward. But what does this mean? Clearly, a balance must be struck.
AI is going to create demands on the attention of voters but will also be used within tools they will rely on to decide which of the candidates best represents their political views. ChatGPT and similar tools are being used by many people to summarise and pre-read emails. It is likely that whatever AI models are available in 2024, they will provide a lens that at least some voters will use to develop their personal view of the election.
Going forward, it should be assumed that it will become increasingly difficult for anyone to really know what is actually real and what is not, when it comes to political information. This may drive candidates to become more focused on in-person events when voters get to see the ‘real deal’ – at rallies and hustings, which arguably may be a good thing.
Artificial intelligence has the capability to help supercharge the messaging of political parties going forward. Reaching greater demographics with clearer and more intrinsic messaging than ever before. But the flip side of this capability is its openness to bad actors utilising AI’s evolution to mislead and misinform.
Political parties should expect greater scrutiny from electoral commissions, as well as the electorate, on their use of AI and the messages it sends, but also on how vocal they are about tackling the threats it presents.
Subscribe to our Editor's weekly newsletter