Earlier this year, thousands of Democratic voters in New Hampshire received a telephone call ahead of the state primary, urging them to stay home rather than vote.

The call supposedly came from none other than President Joe Biden. But the message was a "deepfake". This term covers videos, pictures, or audio made with artificial intelligence (AI) to appear real, when they are not. The fake Biden call is one of the most high profile examples to date of the critical threat that deepfakes could pose to the democratic process during the current UK election and the upcoming US election.

Deepfake adverts impersonating Prime Minister Rishi Sunak have reportedly reached more than 400,000 people on Facebook, while young voters in key election battlegrounds are being recommended fake videos created by political activists.

But there may be help coming from technology that conforms to a set of principles known as "responsible AI". This tech could detect and filter out fakes in much the same way a spam filter does.

Misinformation has long been an issue during election campaigns, with many media outlets now carrying out "fact checking" exercises on the claims made by rival candidates. But rapid developments of AI—and in particular generative AI—mean the line between true and false, fact and fiction has become increasingly blurred.

This can cause devastating consequences, sowing the seeds of distrust in the political process and swaying election outcomes. If this continues unaddressed, we can forget about a free and fair democratic process. Instead, we will be faced with a new era of AI-influenced elections.

Seeds of distrust

One reason for the rampant spread of these deepfakes is the fact that they are inexpensive and easy to create, requiring literally no prior knowledge of artificial intelligence. All you need is a determination to influence the outcome of an election.

Paid advertising can be used to propagate deepfakes and other sources of misinformation. The Online Safety Act may make it mandatory to remove illegal disinformation once it has been identified (regardless of whether it is AI-generated or not).

But by the time that happens, the seed of distrust has already been sown in the minds of voters, corrupting the information they use to form opinions and make decisions.

Removing deepfakes once they have already been seen by thousands of voters is like applying a sticking plaster to a gaping wound—too little, too late. The purpose of any technology or law aimed at tackling deepfakes should be to prevent the harm altogether.

With this in mind, the US has launched an AI taskforce to delve deeper into ways to regulate AI and deepfakes. Meanwhile, India plans to introduce penalties both for those who create deepfakes and other forms of disinformation, and for platforms that spread it.

Alongside this are regulations imposed by tech firms such as Google and Meta, which require politicians to disclose the use of AI in election adverts. Finally, there are technological solutions to the threat of deepfakes. Seven major tech companies—including OpenAI, Amazon, and Google—will incorporate "watermarks" into their AI content to identify deepfakes.

This article is republished from The Conversation under a Creative Commons license. Read the original article.