Around the World, AI-Enabled Deep Fakes Are Poised to Upend Elections

In 2024, voters in at least 64 nations will head to the polls in highly consequential elections. From the U.S. to India, from Venezuela to South Africa, countries comprising close to 50 percent of the world’s population are set to hold national elections this year, the results of which will shape the contours of global affairs for the rest of this decade and beyond.

For all the talk about how this batch of elections will determine how serious the planet is about confronting climate change, or how we will collectively deal with global issues like migration, food security, and global conflict, there is one other reason why these elections will always be remembered: 2024 will be the year when AI-generated deep fakes will alter, and in many cases change, the outcome of democratically-held elections across the globe.

2024, to put it bluntly, will be remembered as the year when fake news and misinformation became full-fledged adults.

Recently, voters in New Hampshire received a robocall purportedly from U.S. President Joe Biden, urging them to abstain from participating in the state’s presidential primary election. The voice, synthesized through artificial intelligence, bore a striking resemblance to the real Joe Biden; it advised recipients to withhold their primary vote, falsely alleging that doing so would impede their ability to partake in the subsequent November general election.

The incident underscored mounting apprehension regarding the ability of generative AI to facilitate the dissemination of misinformation and orchestrate coordinated deep-fake campaigns at scale. In this instance, ultimately the source of the fake calls was uncovered–but that’s here in the U.S. Imagine the havoc that such ploys will wreak on countries with far more fragile institutions and a weaker or nonexistent independent media.

Consider Venezuela—home to arguably the world’s most sophisticated AI-driven misinformation and propaganda machine. The Venezuela system employs an army of thousands of “propagandists” coupled with multilingual deep-fake digital avatars—a lethal combination aimed at upending elections at home with the potential to alter outcomes around the world.

Newsweek spoke to the former President of Colombia, Iván Duque, about the threats that AI-generated synthetic media and fake news will have on global elections this year.

“2024 is the year in history where a record number of people will vote in elections, but this historical milestone will be marked by the influence of social media, AI, and the massification of algorithms,” said Duque. “It is essential for the future of democracy that effective controls are built to identify and prevent misinformation driven by foreign actors to manipulate electoral results. If we do not prevent this, we will all carry the risk of falling into our own artificial democracies.”

Take, for example, the chilling case out of Hong Kong in February, in which a company finance worker was duped by a digitally-created avatar mimicking the company’s chief financial officer (as well as other colleagues that the worker knew in real life) into wiring over $25 million to scammers. With technology that is already so sophisticated that it can trick someone into believing they are speaking with real people they know, imagine the impact that a deep-fake video of a politician who most voters only know through social media and TV could have on swaying an electorate.

To find about what election officials, news organizations, and social media platforms should be doing to confront this existential attack on political news and elections, Newsweek spoke to Louis Steinberg, the former chief technology officer of TD Ameritrade and the founder of CTM Insights, a cybersecurity research lab. “It’s going to be harder and harder to try to detect a fake [so] what news organizations need to do is flip it. Don’t try to detect what’s fake … instead, focus on what’s real.”

Steinberg believes that the key is in creating digital fingerprints for our voices and images, somewhat akin to the facial recognition technology that authenticates our phones and tablets, that can assure the public that certain words were spoken by an individual and in the order that he or she actually said them. “There are new ways to prove that the words that I’m hearing are actually words that came out of your mouth. And you can, you can sign, digitally sign an audio stream,” he said.

Steinberg’s antidote to AI-assisted identity hijacking, albeit a potentially effective safeguard, seems too far off to secure the slate of elections on tap in 2024, which is why he ultimately sees, at least in the near-term, the onus falling back on social media companies. “Social media platforms [need to be] responsible for fact-checking information that they host,” Steinberg said.

Weaponized disinformation has always been a part of politics, but this year AI-generated deep fakes have reached a tipping point, becoming so real and effective that they likely will end up changing election outcomes across the globe.

Arick Wierson is a six-time Emmy Award-winning television producer and served as a senior media and political adviser to former New York City Mayor Michael Bloomberg. He advises corporate clients on communications strategies in the United States, Africa, and Latin America.

The views expressed in this article are the writer’s own.