Campaigns spend millions on AI ads as concerns over propaganda rise
Republican groups spent $137 million on AI-generated political ads by mid-October, sparking debates over psychological targeting and the erosion of shared reality.
The short version
- Republican groups accounted for about 80% of the $137 million spent on AI ads by mid-October.
- AI allows for real-time, automated content creation based on individual psychological data.
- Researchers note that voters often cannot distinguish between human-created and AI-generated content.
- Proposed solutions include mandatory watermarking to flag AI-generated images and content.
Why it matters
AI enables the automated creation of personalized propaganda that can target voters based on their private data. This shift makes it harder for a democracy to function because it removes the common basis of reality required for collective governance.
The rise of AI in campaigns
Political campaigns across the United States have begun integrating artificial intelligence into their advertising strategies. By mid-October, researchers at the Wesleyan Media Project found that campaigns spent more than $137 million on AI-generated ads. This spending represents a significant shift in how political messages are produced and distributed to the public.
Republican groups were the primary drivers of this trend, accounting for about 80% of the total spending on these messages. The content varied widely in style, ranging from surreal humor to more realistic depictions. For example, some ads featured candidates as caped crusaders or gladiators, while others showed more grounded, everyday scenes.
While some ads were clearly intended to be humorous, others were designed to look like authentic photography. One example involved a mailer showing a school board candidate behind a desk, which a critic labeled as a sexist image. As Times of San Diego reported, these diverse uses of AI highlight the technology's versatility in political messaging.
From data to automated propaganda
The use of AI in politics is evolving from the manual data-gathering methods seen in the past. Previously, organizations like Cambridge Analytica had to go through a multi-step process of buying data and predicting psychology to target users. Now, AI systems can potentially create content that targets individuals in real time based on their specific data points.
This capability allows for a more direct and automated form of influence. Instead of just showing a voter a generic ad, AI can theoretically generate content that appeals to a person's unique psychological profile. This process could include using information from credit card records or other private data to influence what a person sees or even their desire to vote.
The concern is that this creates a highly personalized experience for every voter. When content is tailored so specifically to an individual's psychology, it becomes much harder to have a public conversation. Because there is often no disclosure regarding whether a message is AI-generated, the influence can be even more subtle and difficult to track.
The erosion of shared reality
One of the primary dangers of AI-driven political ads is the potential for it to undermine a democracy's foundation. For a democracy to function, people need to share a common basis of reality and truth. However, micro-targeting can create a situation where different people are exposed to entirely different visions of the world.
Research suggests that it is often impossible for voters to tell the difference between human-created content and high-quality AI generation. If voters cannot distinguish between what is real and what is fake, it becomes difficult for them to form a collective agreement on any issue. This creates a fragmented information environment where shared truth is replaced by individual, AI-tailored perspectives.
Experts argue that this is not just a problem of people being fooled by specific deepfakes. It is a broader issue of how data is used to influence what people believe and feel. When machines can use propaganda to influence behavior at scale, it raises questions about whether humans will remain the primary actors in democratic processes.
Regulation and the path forward
To address these concerns, some lawmakers are looking into specific regulations to protect the integrity of elections. One proposed solution is the implementation of mandatory watermarking. A watermark is a permanent tag on an image that indicates it was created by AI, ensuring that people know they are not looking at a human-made photo.
Advocates for these rules argue that regulation is not about stopping innovation but about directing it toward human values. They suggest that AI companies and social media platforms should be held to standards that include disclosing data collection and paying people for their data. The goal is to ensure that technological progress does not come at the expense of social or economic security.
Looking toward the 2028 elections, the focus will likely be on whether leaders can manage these technologies responsibly. If used correctly, AI could serve as a tool to help people explore new ideas and understand the wider world. However, protecting democracy will require active efforts to rein in platforms and ensure that AI serves as a tool for human innovation rather than a replacement for it.
What to watch
- Whether new watermarking laws are passed by state or federal governments.
- How AI-driven micro-targeting affects voter turnout in upcoming local elections.
- The development of new tools that allow voters to verify the authenticity of political images.
Try it
Look for 'AI-generated' labels or watermarks on political ads you see on social media platforms like TikTok or Instagram.
Sources and links
Read the original at Times of San Diego
First reported by Times of San Diego. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.