It’s official. Joe Biden and Donald Trump have secured the necessary delegates^[1] to be their parties’ nominees for president in the 2024 election. Barring unforeseen events, the two will be formally nominated at the party conventions this summer and face off at the ballot box on Nov. 5.

It’s a safe bet that, as in recent elections^[2], this one will play out largely online and feature a potent blend of news and disinformation delivered over social media. New this year are powerful generative artificial intelligence tools such as ChatGPT^[3] and Sora^[4] that make it easier to “flood the zone^[5]” with propaganda and disinformation and produce convincing deepfakes: words coming from the mouths of politicians that they did not actually say and events replaying before our eyes that did not actually happen.

The result is an increased likelihood of voters being deceived and, perhaps as worrisome, a growing sense that you can’t trust anything you see online^[6]. Trump is already taking advantage of the so-called liar’s dividend^[7], the opportunity to discount your actual words and deeds as deepfakes. Trump implied on his Truth Social platform on March 12, 2024, that real videos of him shown by Democratic House members were produced or altered using artificial intelligence^[8].

The Conversation has been covering the latest developments in artificial intelligence that have the potential to undermine democracy. The following is a roundup of some of those articles from our archive.

1. Fake events

The ability to use AI to make convincing fakes is particularly troublesome for producing false evidence of events that never happened. Rochester Institute of Technology computer security researcher Christopher Schwartz^[9] has dubbed these situation deepfakes^[10].

“The basic idea and technology of a situation deepfake are the same as with any other deepfake, but with a bolder ambition: to manipulate a real event or invent one from thin air,” he wrote.

Situation deepfakes could be used to boost or undermine a candidate or suppress voter turnout. If you encounter reports on social media of events that are surprising or extraordinary, try to learn more about them from reliable sources, such as fact-checked news reports, peer-reviewed academic articles or interviews with credentialed experts, Schwartz said. Also, recognize that deepfakes can take advantage of what you are inclined to believe.

Read more: Events that never happened could influence the 2024 presidential election – a cybersecurity researcher explains situation deepfakes^[11]

2. Russia, China and Iran take aim

From the question of what AI-generated disinformation can do follows the question of who has been wielding it. Today’s AI tools put the capacity to produce disinformation in reach for most people, but of particular concern are nations that are adversaries^[12] of the United States and other democracies. In particular, Russia, China and Iran have extensive experience with disinformation campaigns and technology.

“There’s a lot more to running a disinformation campaign than generating content,” wrote security expert and Harvard Kennedy School lecturer Bruce Schneier^[13]. “The hard part is distribution. A propagandist needs a series of fake accounts on which to post, and others to boost it into the mainstream where it can go viral.”

Russia and China have a history of testing disinformation campaigns on smaller countries, according to Schneier. “Countering new disinformation campaigns requires being able to recognize them, and recognizing them requires looking for and cataloging them now,” he wrote.

Read more: AI disinformation is a threat to elections − learning to spot Russian, Chinese and Iranian meddling in other countries can help the US prepare for 2024^[14]

3. Healthy skepticism

But it doesn’t require the resources of shadowy intelligence services in powerful nations to make headlines, as the New Hampshire fake Biden robocall^[15] produced and disseminated by two individuals and aimed at dissuading some voters illustrates. That episode prompted the Federal Communications Commission to ban robocalls that use voices generated^[16] by artificial intelligence.

AI-powered disinformation campaigns are difficult to counter because they can be delivered over different channels, including robocalls, social media, email, text message and websites, which complicates the digital forensics of tracking down the sources of the disinformation, wrote Joan Donovan^[17], a media and disinformation scholar at Boston University.

“In many ways, AI-enhanced disinformation such as the New Hampshire robocall poses the same problems as every other form of disinformation,” Donovan wrote. “People who use AI to disrupt elections are likely to do what they can to hide their tracks, which is why it’s necessary for the public to remain skeptical about claims that do not come from verified sources, such as local TV news or social media accounts of reputable news organizations.”

Read more: FCC bans robocalls using deepfake voice clones − but AI-generated disinformation still looms over elections^[18]

4. A new kind of political machine

AI-powered disinformation campaigns are also difficult to counter because they can include bots – automated social media accounts that pose as real people – and can include online interactions tailored to individuals^[19], potentially over the course of an election and potentially with millions of people.

Harvard political scientist Archon Fung^[20] and legal scholar Lawrence Lessig^[21] described these capabilities and laid out a hypothetical scenario of national political campaigns wielding these powerful tools.

Attempts to block these machines could run afoul of the free speech protections of the First Amendment, according to Fung and Lessig. “One constitutionally safer, if smaller, step, already adopted in part by European internet regulators and in California, is to prohibit bots from passing themselves off as people,” they wrote. “For example, regulation might require that campaign messages come with disclaimers when the content they contain is generated by machines rather than humans.”

Read more: How AI could take over elections – and undermine democracy^[22]

This story is a roundup of articles from The Conversation’s archives.