• AI-generated deepfakes are becoming more common and more convincing, posing a danger for all the elections occurring this year.
• Deepfake videos can be used by bad actors before South Africa’s elections this year to spread misinformation and to discredit political parties.
• When you think you have come across a deepfake, stay calm and verify if the information you see is true.

You simply can’t escape AI these days. The technology already permeates most online spaces, be it social media or otherwise. And while generative AI has a depth of useful capabilities to offer, it can also be abused by internet scoundrels.

A 2024 Global Risks report from the World Economic Forum indicates that AI-generated misinformation is the second biggest global risk, beaten out only by extreme weather conditions. The report takes into account 2024’s electoral climate around the globe, with over 3 billion people expected to vote this year worldwide, including in South Africa come May.

“The widespread use of misinformation and disinformation, and tools to disseminate it, may undermine the legitimacy of newly elected governments. Resulting unrest could range from violent protests and hate crimes to civil confrontation and terrorism,” explains the World Economic Forum.

One of the ways misinformation is spread is through deepfakes. Once obvious and poorly fashioned, current innovations in generative AI have made them nearly impossible to distinguish in many cases. Deepfake videos of politicians spread widely on social media can be used to deceive voters, manipulate public opinion and tarnish the reputation of parties and their members.

“Deepfakes pose major risks in the run-up to elections, both in the UK and South Africa,” says Anna Collard, SVP Content Strategy and evangelist at KnowBe4 AFRICA, a cybersecurity training organisation.

“Voters need to be aware of this risk and be proactive about mitigating their impact.”

Collard shares a notable incident from October last year shows that these incidents are already happening in Africa. Cybercriminals used deepfake videos to impersonate African Union Chairperson Moussa Faki Mahamat during online conversations with European diplomats.

In December last year, Facebook removed over 100 deepfake videos – most paid advertisements – featuring British Prime Minister Rishi Sunak. The videos were already watched by half a million people before they were removed.

“It’s crucial for the public to be aware of the existence of this phenomenon,” Collard emphasises, adding, “The production of deepfake videos and images is cheap and easy.” She shares that entire disinformation campaigns to be spread on social media can be created with a budget of about $400.

According to Collard, politicians should already be anticipating deepfakes made of them for the purpose of spreading fake news.

“If you’re a politician, expect deepfakes impersonating you to appear in order to sway public opinion. Similarly, voters should not believe everything they see or hear on social media to be true.”

Collard explains that before elections, voters must do two things when they believe they have come across a deepfake video on social media: first don’t overreact, and second, verify the information being shared.

“Scammers want to whip up your emotions and get you in a state where you’re unable to think rationally,” she says. “Stay calm and don’t share content that is overly emotive. Whenever your emotions are triggered, slow down and verify, -it could be a hoax.”

“If a politician purportedly said something that’s deeply polarising, it’s important to verify whether they genuinely said it.”

Tools like Padre from the IEC let users fact-check information regarding political parties in the country and their most recent statements. You can also use the likes of Google Reverse Image Search and tools like FotoForensics to find and analyse if images or videos are fake or if they have been tampared with in some way.

However, the truth is that verifying deepfakes now is more difficult than ever.

“Deepfakes are becoming more convincing, making verifying information even harder,” states Collard.

“To combat this form of disinformation and safeguard the democratic process, social media platforms, political parties and independent watchdogs as well as us the public will all need to work together.”