The late Indonesian President Suharto addresses the electorate 16 years after his death and urges them to vote in favour of his party. This is a classic example of deep fake content.

“It’s used to mislead people, to create fake news, fake texts and fake videos. All of this together is then marketed via social media, through fake bots that spread it further,” explains cyber security expert Tomislav Vazdar.

Photos are easier to manipulate, and so the Instagram profile of film director and artist Filip Filkovic combines the incompatible and enters into unimaginable political marriages. It also changes careers. Croatian politicians have also started to improvise with artificial intelligence during the campaign for the upcoming parliamentary elections.

AI could be used to engage people in democratic processes

“I would not expect there to be too many uses at the highest level, like in America, but it’s not impossible. I have heard that certain parties and candidates are preparing for the next local election to use it on a larger scale,” claims Aco Momcilovic, artificial intelligence researcher.

The influence of artificial intelligence will be particularly interesting in the super election year of 2024, when billions of people around the world will go to the polls.

“In America, they have already started with examples. We have fake audio recordings of Joe Biden saying something he never said, fake pictures of Trump taking pictures of himself with people he never took pictures of, and a number of fake texts being published,” says Vazdar.

However, artificial intelligence could also be used for positive purposes to engage people in democratic processes, if desired. Voters may want to simulate a conversation with politicians they sympathise with to learn more about their programmes and opinions.

“Everyone says that politicians should go out and talk to people in person, and this is a technological tool that allows them to be present 24/7 and talk to tens of thousands of people,” says Andrej Steven-Horvat, data consultant.

It is becoming increasingly difficult to recognise fake content

You can not escape artificial intelligence, but you can influence it so that we determine where it gets its data from. The consultant believes that politicians should not be afraid of artificial intelligence.

“They need to engage and feed artificial intelligence with their tested programmes and ideas so that they can be sure that the AI is not picking up publicly available fake news data from the internet that someone has already published,” adds Steven-Horvat.

The biggest problem is that in a digital world dominated by algorithms, it is becoming increasingly difficult to recognise fake content. They lock us into the bubbles of like-minded people and radicalise us more and more. And it could even get worse.

“On top of what we are already seeing today, there’s a new level of artificial intelligence that’s going to amplify it. Because there is even more information about each individual person that can be used to create customised messages to which they respond best,” says Momcilovic.

Whether politicians will be able to respond appropriately to these challenges or whether they will also try to misuse artificial intelligence is a question that raises many fears.