In the rapidly evolving landscape of digital advertising, concerns surrounding artificial intelligence (AI) generated content, especially in political ads, are prompting major platforms to take action. The emergence of sophisticated AI tools has led to a new kind of content, often termed 'slopaganda,' which blends real and fabricated visuals to sway public opinion.
YouTube's Approach to AI Content
YouTube has implemented specific measures to tackle AI-generated content in political advertising. According to a spokesperson, the platform enforces prominent labels on realistic AI-generated material to keep viewers informed.
For political candidates, YouTube allows them to request the removal of content that features altered or AI-generated representations of their likeness. The platform also mandates that advertisers disclose when synthetic or digitally manipulated people or events appear in election advertisements.
Google's Verification Requirements
YouTube's parent company, Google, has a stringent verification process for political ads. All participating advertisers must complete this process and include necessary disclosure statements indicating who paid for the advertisement. This ensures transparency and accountability in political campaigning.
Social Media Guidelines and Recent Developments
On the other hand, X (previously Twitter) places the responsibility for compliance with applicable laws squarely on advertisers, not the platform itself. The company did not provide a comment on its current policies regarding AI-generated ads.
Meta, another major player in social media, has outlined its guidelines for political advertisements, especially concerning their use within a week of election dates. They also require the "paid for by" disclosures and maintain an extensive ad library for transparency's sake. In June, Meta enhanced its labeling system by introducing an “about this ad” tool for advertisements suspected of being created using third-party AI tools.
However, Meta has faced significant legal challenges, including a recent $25 billion penalty ordered by a New Mexico court related to data privacy violations stemming from the infamous Cambridge Analytica scandal. This scandal involved the improper access of data from 87 million Facebook users by a political consulting firm.
Challenges with Generative AI
The rise of generative AI also presents unique challenges for tech platforms. According to a study conducted by the Brennan Center for Justice, various AI models—such as ChatGPT and others—were tested for their propensity to create misleading content. The findings revealed that while some AI models could be manipulated into generating deceptive material, they also showed improvements, like redirecting harmful prompts to non-deceptive outputs, such as illustrations.
Legislative Responses to AI in Political Advertising
In response to these challenges, several lawmakers are taking proactive steps. The bipartisan Protect Elections from Deceptive AI Act proposes prohibiting the distribution of materially deceptive AI-generated video or audio related to federal candidates. Additionally, other initiatives, like the AI Ads Act led by Democratic representatives, aim to expand prohibitions on campaign misrepresentation to include content generated by AI.
Despite these efforts, none of the proposed bills have yet advanced beyond committee discussions, highlighting the complexities of regulating AI in political advertising.
