Malta enters its first AI-era general election as truth comes under pressure

Malta’s first general election in the age of artificial intelligence is already reshaping political campaigning, with parties turning to AI-generated visuals and videos to sell their vision for the country.

The Nationalist Party has used AI-generated artistic impressions to visualise some of its proposals, including the planned National Hall forming part of its proposed Culture District. But the images also highlighted one of the growing concerns surrounding AI-generated political material.

The “AI-generated idea of the hall”, as one PN spokesperson described it, appeared to closely resemble an existing proposal for Helsinki’s new design museum found online.

Meanwhile, the Labour Party has published at least three AI-generated claymation-style videos, narrated by what appears to be a generic AI-generated American voice, to explain several of its proposals for Malta’s future. Also, press releases and commentaries are being entirely generated by AI.

The use of AI by both parties reflects how Malta’s political battlefield has moved beyond house visits and coffee mornings to the phones in voters’ hands.

Experts warn that while AI can help voters analyse information more efficiently, it can also expose them to misleading or entirely fabricated content at an unprecedented scale.

Against this backdrop, there are increasing concerns about Malta’s particularly heavy reliance on social media for news.

Malta has the highest proportion of people in the EU who rely on social media as their main source of news. Around 85 per cent of respondents in Malta said they use social media for news, compared with an EU average of 74 per cent, according to Eurobarometer data.

An example of an AI generated image of Alex Borg that appeared just as the election was announced

“The biggest concern is that people do not know what is real, are exposed to this technology every day and become hooked on it,” said Mark Bugeja, a lecturer in the University of Malta’s Department of Tourism Management with an academic background in AI.

“There is a concern that people could be dissuaded from voting because they see something fake that does not align with their views.”

To better understand how AI can be used responsibly during the election, and how voters can avoid being misled, Times of Malta spoke to Bugeja and Dylan Seychell, a lecturer at the University of Malta’s Department of Artificial Intelligence, soon after the election date was announced.

The best way to tackle misinformation and disinformation, for now, is to rely on personal judgement, Bugeja said.

“Do not trust everything you see. Ask yourself if it makes sense, ask why, and do your research,” he said.

Misinformation refers to false or inaccurate information where facts are wrong but not necessarily intended to mislead. Disinformation, on the other hand, is deliberately false information designed to deceive.

Bugeja said one of the most effective ways to protect oneself is to consult multiple sources before accepting a claim as true.

“Even before AI, if you saw a news item and just one outlet was reporting it, you would doubt whether it was true. But if everyone else is reporting it, then you know it is probably true,” he said.

He stressed that the source of a news item is crucial, adding that readers should not only check whether other outlets have reported the story but also assess whether the source itself is credible.

For example, a Facebook page named Infinite Facts amassed around 23,000 followers by posting feel-good stories, many of which were later found to be fake. The page has since disappeared and could re-emerge elsewhere under a different name.

“You cannot just see a picture on social media and assume it’s true,” Bugeja said.

Fake images can be difficult to detect, especially when they do not include people. But when people are shown, tell-tale signs may include distorted hands, unnatural lighting or shadows that do not make sense.

‘Everyone is being bombarded’

Seychell, a principal investigator on AI literacy and involved in a national AI literacy initiative led by the Malta Digital Innovation Authority and the University of Malta, said his main concern is the sheer volume of information now being produced.

“Everyone is being bombarded with information and could have information that is misleading and not even notice,” he said.

At the same time, Seychell noted that AI can also be used to help individuals navigate complex information.

“You can fight fire with fire and use AI to your benefit, such as by analysing long and sometimes tedious data,” he said.

“When both parties release their manifestos, you can ask AI to analyse and compare them for you. Let us say you tell it you are a student from Gozo. It can give you a personalised debrief on how each one affects you positively or negatively.”

He cautioned, however, that AI systems are not infallible and can make mistakes, meaning users must still apply critical thinking when interpreting results.

AI has also been widely adopted by voters and supporters themselves. Following the release of billboards by the Labour Party and the Nationalist Party, users created numerous memes and satirical images based on them. Others used AI to generate images of themselves as candidates for their respective party.

While some were light-hearted, others included xenophobic and racist AI-generated images targeting Malta’s Indian community.

Internationally, there have already been several instances where AI-generated content influenced election campaigns.

The Global Investigative Journalism Network reported that an AI-synthesised impersonation of the voice of an opposition leader helped swing the election in Slovakia to a pro-Russia candidate. The network also documented the use of AI-generated audio in elections in countries from the US to India to Ethiopia.

“We need to address our literacy,” Seychell said.

“If people were willing to believe a scam and give their own money away, can you imagine what they could do with their vote?”