For years, online advertising has been built on a simple idea: if companies want to show people relevant ads, they need to collect large amounts of personal data.
This often means tracking users across websites and apps to learn about their interests and behaviors.
But new research from the University of Kansas suggests that this trade-off may not be necessary. Artificial intelligence may be able to deliver useful and relevant ads without spying on users.
The study was led by Vaibhav Diwanji, an assistant professor of journalism and mass communications at the University of Kansas.
His research explored whether AI can understand the context of what people are reading or viewing online and use that information to deliver appropriate advertisements without collecting personal information.
According to Diwanji, the digital advertising industry has spent the past two decades assuming that better targeting requires more surveillance.
Every improvement in advertising has generally been linked to collecting more user data and tracking people’s online activities.
However, the new research suggests there is another way.
Instead of building a profile of a user, AI can focus on the content that the person is currently viewing. By analyzing the structure, meaning, and emotional tone of a webpage, AI can determine what types of advertisements might be relevant at that moment.
To study this idea, Diwanji conducted four experiments involving more than 1,000 participants. In each experiment, people visited websites that displayed AI-generated contextual advertisements and then answered questions about their experiences.
The first study found that animated advertisements attracted more attention than static ones. Participants considered the moving ads more valuable, had more positive opinions of the brands, and were more likely to consider making a purchase.
The second study looked at where advertisements appeared on a webpage. Ads placed directly within articles performed better than those placed on the side of the page or in separate sections. People noticed these ads more often and did not view them as especially disruptive.
The third study examined whether advertisements matched the content of the webpage. When ads were closely related to what people were reading, participants paid more attention to them and processed the information more easily. This led to higher perceived value and better engagement.
The final study investigated whether the type of product being advertised mattered. Products that typically require more thought before purchasing, such as automobiles, health insurance, or cruise vacations, generated stronger responses than everyday products like candy or cleaning supplies.
Taken together, the studies suggest that AI can create personalized experiences without actually knowing anything personal about the user.
This finding comes at an important time. Governments and regulators around the world have become increasingly concerned about online privacy and the widespread collection of personal data. Laws in Europe and California have already introduced stricter rules on digital tracking, and the technology industry has been moving away from third-party tracking cookies.
The new research suggests that effective digital advertising and privacy protection do not have to be opposing goals.
Instead of following people around the internet, AI may be able to understand what matters to users simply by understanding the content they are engaging with in the moment. This approach could create a future where online experiences remain relevant while better respecting personal privacy.
