AI Experts Give Mixed Warnings On 2024 Election Impact

(BrightPress.org) – Artificial intelligence experts are concerned that the ability to generate and spread misinformation rapidly could interfere with the 2024 election. Professor emeritus of cognitive science at NYU Gary Marcus told Fortune magazine “We should be scared sh**less.” His comments follow calls from Elon Musk and Apple CEO Steve Wozniak to pause advanced AI development as the systems become increasingly advanced. 

As recently as March, fake AI-generated arrest photos of Donald Trump were circulating in an attempt to troll or test public credulity. Many quickly realized they were fake, but the problem of AI-generated misleading information remains.

An AI-generated recording of Chicago mayoral candidate Paul Vallas made the moderate Democrat who was endorsed by the Chicago police union seem like he had no problem with cops killing suspects. Again, the fraud was detected early and most don’t believe it impacted the extremely close run-off election, but the problem presented by fake audio remains. 

Marcus warned of the potential impact of foreign adversaries to mislead American voters but was relatively quiet on the possibility of American billionaires using the technology to push the needle toward their desired political candidate. He also warned of the potential for a “firehose of falsehood” that is a non-stop flood of deception that could overwhelm social media companies’ ability to react. 

Chris Meserole is a fellow at the Brookings Institute, and he specializes in AI. He seemed to think the potential impact is exaggerated as the technology isn’t yet advanced enough to cause serious problems. He did express concern that there could be a pivotal moment when some AI-generated content could surface that would be difficult to prove was faked.

The problem may even be more subtle. A German study found that people are likely to underestimate the impact of statements from AI chatbots, even when they know it’s an AI making the statement. With the advent of more advanced technologies that allow AIs to effectively mimic humans on social media, how will anyone be able to parse genuine human opinion from AI-generated manipulative spam?

Copyright 2023, BrightPress.org