AI, Threat to Elections and Democracy

Did Joe Biden really call voters to discourage them from voting in the New Hampshire primary last January? No, it was a “deepfake” audio, which was only one of the many cyber-threats hanging over elections, warns Olivier Godin, Zscaler’s Managing Director for Southern Europe.

In terms of elections, 2024 is an important year. As half the globe is called to vote, there have never been so many cyberattacks fueled by artificial intelligence.

According to the latest global risks report from the World Economic Forum, we should take misinformation and false information very seriously.

The Inevitable Increase in Deepfakes

The growing use of deepfakes is at the heart of the matter, and the more mature the technology behind these deepfake audios and videos becomes, the more we must expect to see new examples flooding social media.

Cases earlier this year include the “robocall” that used Biden’s voice to dissuade voters from voting in the New Hampshire primary in January 2024. Similar tactics, although more malicious, have also been employed in Slovakia and the U.K.

Until recently, deepfakes have focused on public figures, but they could now target election officials in highly contested constituencies. Fake websites are used to spread false information and, being accessible online, they can be shared quickly, which increases the impact of the deepfakes.

Threat of Hacking

It was because of these risks that, last February, hundreds of leading figures in the artificial intelligence community signed a letter demanding criminal sanctions for those creating and sharing harmful content. However, even if these new rules are enacted in time for the election, it is unlikely that it would change anything.

Attackers could choose to delay sharing this false information until the most opportune moment in the election calendar. Some attacks that have gone undetected for months, even years, could cause serious leaks.

People blamed such election meddling during the 2020 U.S. presidential election mostly on Russia, and this tendency could continue in 2024. However, other countries, political groups and individuals could also use their resources to influence voters or disrupt the process. Recent reports explain how China intervened in the Canadian federal elections in 2019 and 2021. There is also evidence that the Chinese are willing to disrupt the 2024 U.S. election.

Targeted Election Infrastructure

Voting machines are another prime target for state-sponsored hackers. If the security of these machines were to be compromised in the next U.S. election, it would only increase charges of election fraud conspiracy like those promulgated by former President Donald Trump in the days after his defeat in 2020. Unlike last time, these attacks could then be proven.

The Cybersecurity and Infrastructure Security Agency is prepared for these attacks. The website #protect2024 offers a wave of content centered around security for state and local election officials. These resources are designed to reinforce security hygiene practices, improve the level of system security, and simplify incident management.

‘Fake Websites Are Used To Spread False Information.’

The role of ethical hackers is also essential for examining the vulnerabilities of material and software used by election technology manufacturers.

Election-related security issues are a good illustration of the broader challenges affecting all business sectors. We will need to learn from the incidents affecting this year’s elections and put measures in place until we can use artificial intelligence to counter attacks before they cause irreversible damage to our democracies.