Rapid recent advances in artificial intelligence (AI) systems are leading to increasing concern that the output of the technology will be misused during forthcoming elections.
In 2024, approximately half the world's adult population will be asked to vote.
Over the last twelve months, several AI companies have released software and tools which make it possible to easily create images, audio and video which are almost indistinguishable from authentic versions of the same content.
As a result, regulators and some companies are scrambling to find ways to limit the use of AI-generated content for misinformation or disinformation purposes.
With local elections, European Parliament elections and a general election all due to be held in Ireland soon, Katie Hannon spoke to Professor Barry O’Sullivan of University College Cork for 'Upfront: The Podcast’ about the recent development of AI and the threat it could pose to democracies across the world.
"I really believe we're going to see lots of content in elections internationally that's going to really be impacting the psychology and the agency of voters, and that will change the outcome of elections," Professor O'Sullivan told Katie Hannon.
"In 2025, we will be seeing reports – if not earlier – of election results where there is significant evidence that the outcome has been impacted by the use of AI systems."
"That is a terrifying prospect, that’s why I think that's why we need a call to arms internationally," Professor O’Sullivan, who is a past President of the European AI Association, said.
"The very fabric of our society is based on democracy and if we’re not certain we’re getting the true expression of the electorate that would be a pretty serious situation," he added.
LISTEN: Professor Barry O'Sullivan speaks to Katie Hannon on Upfront: The Podcast
We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences
Since October 2023 attempts have been made to influence voters using AI-generated video and audio during elections in Slovakia, Pakistan and the State of New Hampshire in the United States.
"I tell people they should really be looking at anything that doesn't come from a trusted media source with a degree of scepticism. They should try to at least ask some basic questions on the provenance of it. 'Who created it? Where did it come from? How do I know it's real? Could it be fake?' And to ask themselves that question. If something isn't being corroborated from multiple sources, the chances are it is fake."
Several of the world’s leading AI companies have their European headquarters in Dublin, and Professor O’Sullivan says as a result the Irish government should be more proactive to address concerns.
"From an Irish point of view, given that we host these companies, I think with that benefit - with that honour - comes responsibility," Professor O’Sullivan told Katie Hannon.
"I would much prefer that we are proactive in ensuring [AI-related disinformation] does not occur, rather than discovering in twelve months' time that it did actually happen, when we didn't think it would."
Last week, several major technology companies signed an accord in Munich agreeing to try to limit the spread of AI-generated misinformation. Professor O’Sullivan says - while welcoming such steps - that more measures by be required.
"Ireland could really play a significant leadership role in creating... a task force. And that task force would be doing multiple things. It would be ensuring these companies are sharing information, that they've established and that they've created some sort of best practice to deal with this, that they're going to going out to the public and helping inform the public about what to look for, making this information accessible."
Asked if the companies themselves are doing enough to tackle AI-related issues, he said "my sense is no they’re not."
"One of the reasons I believe that is that last year we had all of these companies going to the US Congress and begging to be regulated because they were going out of control."
"I would have thought that if they were serious then, they would have maybe plugged out some of these systems until they were convinced that they were safe. But they didn't do that."
Professor O’Sullivan last year became the first Irish person to win the European Association for Artificial Intelligence’s EurAI Distinguished Service Award.
He has advised government agencies, companies and NGOs on AI internationally.
"This year the big one is the potential use of AI to undermine the outcome of elections, that's something we really need to take very seriously," Professor O’Sullivan said.
"I think everything we do will help. We might not eradicate [AI-related disinformation]. I don't think we will... but anything that's done to sort of chip away at the risk I think is going to be beneficial."
Listen to Professor Barry O'Sullivan speaking to Katie Hannon on Upfront: The Podcast here, on Apple Podcasts and on Spotify.