Misinformed on Misinformation: Why Generative AI Won’t Harm Democracy in 2024

Despite not being mentioned in the presidential debate on June 27th, generative AI is at the forefront of voters’ mind in the upcoming election. With the assassination attempt and Biden exiting the race, it’s easy for generative AI to take a backseat as an issue. One current concern with generative AI is the creation of fake news. Although fake news claims have increased in the past 10 years, viewership of untrustworthy sources has decreased (Witte).

Why Fear GenAI?

Fears of fake news written by generative AI boil down to two categories: pervasiveness and persuasiveness.?

With the relatively new ability to generate entire articles with the press of a button, fears of fake news flooding the market skyrocketed. Not only would more misinformation increase the total spread, but the logic follows that misinformation would drown out factual reporting (Verma).

Meanwhile, generative AI also makes it easier than ever to produce sophisticated content. Unlike past attempts by foreign governments to spread misinformation, which were plagued with spelling mistakes and grammatical errors, generative AI is able to produce cohesive sentences, and use high-level language, thus improving the credibility to an uninformed reader (Metzger).

Insight About Fake News

“While concerns about the potential impact of generative AI on misinformation are valid, they are often exaggerated and not supported by empirical evidence” (Simon et al.) .

It’s no secret that generative AI is changing and will change everything. That said, the outlook shared by 53% of Americans that AI will alter the outcome of the upcoming election is worrying (“National Tracking Poll”).

Although generative AI will produce a record amount of fake news campaigns online, the impact of those campaigns in comparison to the current output of fake news will not be proportional. The assumption that an increase in fake news directly correlates to a spike in views is predicated on the belief that the misinformation is inherently un-saturated.?

In terms of quality of production, online tools have made it simple for peddlers of fake news to create professional content for years. Yet, professional-seeming untrustworthy articles are often swept under the rug, as most fake news flies under the radar (Simon et al.).?

Despite a massive increase in fake news articles and their quality, it’s unlikely that major media outlets will ever be drowned out; 96.54% of all information consumed by Americans between 2017-2021 was trustworthy (Altay et al.). Even on social media platforms, which are suspected to be the primary purveyors of fake news, 89% of all views on news were from trusted sources (Guess et al.). Lastly, as there is some reassurance in the fact that in prior election cycles, the effect of campaigns and information itself during election cycles has been nearly zero (Kalla and Broockman).

However, predicting the exact effect of AI-generated news is made difficult by the constantly updating frameworks. Because of the massive steps taken in the past 2 years, there is a clear lack of rigorous academic media literature specifically about generative AI.

The Real Danger of GenAI Fake News

While there will be a small, tangible effect, AI generated fake news is unlikely to, in its current form, affect the outcome of the 2024 election. Yet, there remains an inherent danger, and it lies in an erosion of trust.

Fake news carries almost no power unless major media institutions are torn down. Just a small cohort of “superspreaders” view and share 80% of all fake news. The common denominator among these superspreaders is simple: a lack of trust in institutions (Grinberg). Deteriorating public trust in the media and stripping down institutions only serve to exacerbate the issue of fake news.

Renée DiResta, the former research director of the Stanford Internet Observatory, which was recently closed after a slew of lawsuits, said, “The work of studying election delegitimization and supporting election officials is more important than ever. It is crucial that we not only stand resolute but speak out forcefully against intimidation tactics intended to silence us and discredit academic research. We cannot allow fear to undermine our commitment to safeguarding the democratic process” (DiResta).

Even before the release of ChatGPT, it was clear that AI would progress faster than anything we’d ever seen, barring one roadblock: trust (McKendrick).

It’s no secret that a majority of Americans are fearful of the future of AI (Douglas)(Rampe). That said, most of the public’s fears come without proper knowledge of the tools that could “destroy humanity” (Dwork).

All in all, it’s vital that AI experts share information about the importance of supporting advancement in AI research. Unless action is taken, the diffusion of misinformation using AI tools, and further erosion of public trust could halt future progress. In its current state, the public’s concerns about generative AI will, over time, outweigh the costs of the fake news itself.