Understanding why do People engage with Unreliable Websites
In today’s digital landscape, the spread of misinformation is a growing concern. Many discussions have centered around how people come across unreliable information online, especially through search engines. As part of a team of researchers from 美国普林斯顿大学 and 微软 AI For Good Lab, we co-authored a recent study that takes a close look at this issue. Our findings reveal something important: most of the time, when people engage with unreliable websites from search it’s because they actively seek them out, not because they are unknowingly led there.
The Study: What Were We Trying to Understand?
Our team analyzed approximately 14 billion search results from Bing to explore the relationship between user behavior and engagement with unreliable websites. The key question we wanted to answer was: Are search engines directing users to unreliable information, or are users making deliberate choices to visit these sites?
We examined search behaviors, focusing on both the results search engines present and what users actually click on. We used ratings from trusted fact-checking organizations, such as NewsGuard, to classify websites as reliable or unreliable, and tracked how often users interacted with them.
Key Finding: It’s About What Users Seek, Not What Algorithms Push
Our study found that the vast majority of engagement with unreliable websites from search happens when users are explicitly searching for specific sites. More than 80% of visits to unreliable websites were a result of users deliberately searching for them by name. In other words, most of the time people visit these sites from search they are doing so by choice, not by accident.
Interestingly, we found that only a tiny portion—less than 2%—of all searches were aimed at unreliable sites, but these searches accounted for vast majority of engagement with these spaces. ?This highlights the significant role that individual user preferences play in the spread of unreliable information.
User Preferences Drive Engagement
One of the most revealing aspects of our research was how user preferences influence online behavior. Many users employ “navigational searches,” where they type in the name of a specific website to visit it directly. For example, someone might search for “Natural News,” a site that has been flagged for promoting health misinformation.?
This means that search engines like Bing are not pushing unreliable sites to unsuspecting users. Instead, engagement with unreliable sites occurs largely because users are looking for them.
领英推荐
Implications for Combating Misinformation
Our research offers valuable insights into how we can address the challenges from dis- and misinformation online. It’s not just about adjusting algorithms; it’s also about understanding why people are seeking out unreliable information sources in the first place.
Here are a few strategies that could help:
Final Thoughts
Our research shows that it’s not simply algorithms steering users toward low-quality sources—it’s largely driven by what people are searching for.
This has important implications for how we address challenges in the information environment. It suggests the real key is understanding user behavior and preferences. By combining smarter search solutions with media literacy initiatives, we can help create a more informed and responsible online environment.
Our study adds an important piece to the puzzle of how people engage with information online and brings us one step closer to building a more trustworthy digital ecosystem.
"Secure GenAI" Newsletter | Book author & technologist.
3 周This week my newsletter also highlight similar concern, especially Iranian Cyber Playbook. More: https://lnkd.in/gJDW9xBS
AI Team Lead | PhD in AI
3 周Can the search engines also detect and highlight if a website is unreliable? What resources would it require?
Product Manager | Ph.D., Biostatistics | Data Science | Machine Learning | Transformative AI products and AI for Good
3 周Yes, people believe these lies because they want to and that is the real question. Why?