How AI-generated ads can 'backfire' on politicians
Oct 10, 2024
auto_awesome
Scott Babwah Brennen, Director of the Center on Technology Policy at NYU, leads research on AI's role in political advertising. He dives into a striking study where AI disclaimers in political ads surprisingly undermine candidates' credibility. The conversation explores how these ads challenge public trust and the implications of misleading information. Brennen advocates for better regulations to improve transparency and labeling, highlighting AI's subtle yet impactful effects on voter perception and political engagement.
AI disclaimers in political ads often decrease voter trust in candidates, leading to a potential backfire effect on their public image.
The confusion surrounding generative AI usage in political advertising highlights the need for better communication strategies to inform voters on authenticity.
Deep dives
Impact of AI Labels on Candidate Perception
AI labels in political ads significantly affect voter perception of candidates. When voters see an ad that discloses AI involvement, their trust in the candidate who created the ad typically decreases. For instance, those who viewed attack ads with an AI label rated the candidate’s trustworthiness and appeal lower compared to candidates without such labels. This suggests that AI disclaimers may inadvertently become detrimental to a candidate's image, illustrating a potential 'backfire effect.'
Challenges of AI Disclosure in Political Advertising
The effectiveness of AI labels in disclosing advertisement authenticity appears limited, as many respondents did not identify AI as the technology used in ad creation. A study indicated that without an explicit label, voters often perceived the ad’s content as resulting from traditional video editing rather than generative AI. This underlines the confusion surrounding generative AI in political contexts, which can lead to voters being misinformed about the nature of the content they consume. The research indicates the need for clearer communication strategies regarding AI usage in political advertisements.
Partisanship and the Effects of AI Labels
Partisanship significantly influences how voters respond to AI-labeled political ads, with effects primarily seen among individuals assessing ads from their own party. Democrats tended to lower their opinion of fellow Democratic candidates using AI, while Republican views on Democratic candidates remained unaffected. This behavior suggests a pre-existing skepticism towards candidates of their own affiliation, raising concerns about how AI might amplify distrust in political messaging. Consequently, this phenomenon could hinder mobilization efforts and fundraising within political campaigns, as voters may be less likely to support candidates negatively impacted by AI labels.
A recent experiment involving two fake political ads tested voter reactions to seeing AI disclaimers — with surprising results. The study’s lead researcher Scott Babwah Brennen, the director of the Center on Technology Policy at NYU, joins Steven Overly to break down the findings.