AI is making it harder to trust your own eyes this election

Political campaigns have always struggled with credibility issues. As artificial intelligence becomes more prevalent in political advertising, it may become even more difficult for voters to trust what they see and hear. Recent examples of AI-generated campaign ads have raised concerns about the potential erosion of public trust in politicians and institutions. Experts warn that the increasing use of AI in political media could deepen the erosion of trust that has been occurring in recent years.

Nate Persily, co-director of the Stanford Law AI Initiative, believes that Americans are losing faith in elite institutions to discern truth from falsehood, and that AI is exacerbating this trend. He argues that the more artificial media people consume, the less likely they are to distinguish between truth and falsehood, and the less likely they may be to care about the information they receive.

A recent example of AI-generated political advertising comes from Texas Senate nominee Ken Paxton. He recently released a campaign ad that featured a synthetic version of Democratic nominee James Talarico delivering some of his most controversial statements. The ad used AI to generate Talarico's voice and facial expressions, while manipulating his actual remarks about topics such as biological sex, God, and Christianity.

The ad was based on real statements made by Talarico, but the synthetic version of him delivering those statements altered their context.

The use of AI in political advertising has sparked a broader backlash against the technology, with both conservatives and liberals expressing a lack of trust in AI-generated media. Sarah Kreps, director of Cornell's Tech Policy Institute, notes that people want authenticity and not AI-generated content. Persily suggests that politicians may be better served by simply presenting their own quotes without the risk of blending them with AI-generated content.

The main concern with AI-generated political ads is not that people will be deceived by the content, but that bad actors can use AI to dismiss real statements as deepfakes. Even when voters recognize an ad as synthetic, the emotional impact of the content may persist, leading to negative feelings about the politician involved.

While not all political AI is equally deceptive, the lack of clear federal legislation leaves a patchwork of state rules that vary in their requirements for disclosing AI-generated content in political ads. Some states require disclosures for potentially deceptive AI-generated ads, while others impose looser restrictions on clearly satirical content.

In New York, for example, state Democrats recently asked election officials to investigate a Republican gubernatorial campaign's AI ad for allegedly missing material deceptive labels, but the Republican claimed the ad was satire.

Without clearer rules and regulations, voters may face a confusing landscape of AI campaign standards that could push them away from shared facts and further erode trust in politicians.