It is official. Joe Biden and Donald Trump did it secured the obligatory delegates to be their parties' candidates for president within the 2024 election. Barring any unexpected events, the 2 shall be formally nominated on the party conventions this summer and can face off on the ballot box on November fifth.
It's secure to assume that this, just like the recent election, shall be largely online, with a potent mix of reports and disinformation across social media. New this 12 months are powerful generative artificial intelligence tools like ChatGPT And Sora that make it easier, “flood the zone” with propaganda and disinformation and produce convincing deepfakes: words from the mouths of politicians that they didn’t actually say and events unfolding before our eyes that didn’t actually occur.
The result’s an increased likelihood that voters are being deceived and, perhaps just as worryingly, a growing sense that that is the case You can't trust anything you see online. Trump is already exploiting the so-called Liar's dividend, the power to dismiss your actual words and actions as deepfakes. Trump suggested on his Truth Social platform on March 12, 2024 that these were real videos of him shown by Democratic House members be created or modified using artificial intelligence.
The conversation covered the newest developments in artificial intelligence which have the potential to undermine democracy. Below is a summary of a few of these articles from our archive.
1. Fake events
The ability to create convincing fakes using AI is especially difficult in the case of providing false evidence of events that never happened. Computer security researcher on the Rochester Institute of Technology Christopher Schwartz has called this case deepfakes.
“The basic idea and technology of a situational deepfake are the identical as every other deepfake, but with a bolder goal: to govern an actual event or to invent one out of thin air,” he wrote.
Situational deepfakes may very well be used to spice up or undermine a candidate or suppress voter turnout. If you come across reports of peculiar or extraordinary events on social media, attempt to learn more about them from reliable sources, corresponding to fact-checked news reports, peer-reviewed scientific articles or interviews with recognized experts, Schwartz said. Also, recognize that deepfakes can exploit what you are likely to imagine.
2. Target Russia, China and Iran
The query of what AI-generated disinformation can achieve results in the query of who used it. Today's AI tools enable most individuals to provide disinformation. Of particular concern, nevertheless, are countries which can be opponents of the United States and other democracies. Russia, China and Iran particularly have extensive experience with disinformation campaigns and technology.
“Running a disinformation campaign is about way more than simply generating content,” wrote the safety expert and Harvard Kennedy School lecturer Bruce Schneider. “The hard part is distribution. A propagandist needs a series of faux accounts to post on and others to bring the thing into the mainstream where it may go viral.”
According to Schneier, Russia and China have long been testing disinformation campaigns against smaller countries. “To counter latest disinformation campaigns, you want to find a way to detect them, and to detect them, you now have to look for and catalog them,” he wrote.
3. Healthy skepticism
But it doesn't require the resources of shady intelligence agencies in powerful countries to make headlines, because it does in New Hampshire fake Biden robocall created and distributed by two people and illustrated with the aim of deterring some voters. This episode prompted the Federal Communications Commission to ban robocalls that use voices generated by artificial intelligence.
AI-powered disinformation campaigns are difficult to combat because they could be delivered through multiple channels, including robocalls, social media, email, text messages and web sites, complicating digital forensics to trace down the sources of disinformation, he wrote John Donovana media and disinformation scholar at Boston University.
“In some ways, AI-powered disinformation just like the New Hampshire robocall raises the identical problems as every other type of disinformation,” Donovan wrote. “People who use AI to disrupt elections are likely doing their best to cover their tracks. Therefore, the general public must remain skeptical of claims that don’t come from verified sources, corresponding to local television news or social media accounts from reputable news organizations.”
4. A brand new sort of political machine
AI-powered disinformation campaigns are also difficult to combat because they will involve bots – automated social media accounts posing as real people – and online interactions tailored to individuals, potentially over the course of an election and potentially with hundreds of thousands of individuals.
Harvard political scientist Archon Fung and legal scholar Lawrence Lessig described these capabilities and designed a hypothetical scenario of national political campaigns using these powerful tools.
According to Fung and Lessig, attempts to dam these machines could violate the First Amendment's free speech protections. “A constitutionally safer, albeit smaller, step, already adopted partially by European web regulators and in California, is to ban bots from impersonating humans,” they wrote. “For example, regulation could require campaign messages to contain disclaimers if the content they contain is generated by machines relatively than humans.”
Disinformation is rampant on social media – a social psychologist explains the tactics used against you
Misinformation, disinformation and hoaxes: what's the difference?
Disinformation campaigns are a dark mixture of truth, lies and sincere belief – lessons from the pandemic
