You open social media without a clear purpose. Maybe you are waiting for coffee, avoiding an email, or trying to escape the room you are sitting in. You are not planning to read the news. You are simply scanning whatever appears in the feed.

Then a headline stops you.

For someone in their twenties, it might be a story about graduates using AI tools to prepare for job interviews. For someone in their thirties, it might be housing pressure, career change, fertility technology, or chatbots in office work. For someone in their forties, it might be children’s education, parents’ healthcare, retirement planning, or the future of work.

You did not search for this topic. You do not read about it every day. Yet it feels strangely close. It touches something you have been thinking about in the background, perhaps without giving it a clear name. You click because the headline seems to say: this is one of your interests too.

As a researcher, I am interested in this small but powerful moment. This article, “Surprise Me: Serendipitous News Discovery and News Consumption,” coauthored with Dr. Erin Ruppel, asks what happens after news unexpectedly finds us. Why do some accidental encounters disappear from attention, while others make us want to keep reading?

When serendipity enters the news feed

Serendipity studies often begin with a familiar human experience: we sometimes find valuable things without intentionally looking for them. A wrong turn in a library, a conversation we overhear, or a link we nearly ignore can lead us to something useful. Research ideas often emerge this way as well: not from a perfect path, but from noticing an unexpected connection.

My work brings this idea into algorithmic news environments. Many people encounter news through recommender systems, including social media, news aggregators, and web portals. These systems learn from what we click, read, like, and share. They help us find relevant information quickly, but they may also keep showing us similar content. Over time, people may feel trapped in a loop of familiar topics.

In our paper, serendipitous news discovery means encountering news that is both unexpected and useful. It is not random content. Randomness may surprise us, but it often has no value. It is also not perfect personalization. Perfectly personalized news may be useful, but it can become predictable. Serendipitous news sits in the middle: it introduces something we were not looking for, but somehow still wanted to know.

Surprise is the spark

The key insight of our study is that surprise helps explain why serendipity matters for news consumption.

In our online experiment, participants in the United States first selected a news topic they were interested in, such as climate change or genetically modified organisms. They then read news in a recommender-system-like environment and selected one recommended article to continue reading. Afterward, they reported whether they had experienced serendipitous discovery, how surprised they felt, and whether they wanted to continue consuming news.

The results showed a clear pattern. People who experienced more serendipitous news discovery felt more surprised. That surprise was then associated with stronger intentions to keep consuming news. Surprise worked like a bridge. It helped transform an accidental encounter into continued exploration.

This surprise is not shock, clickbait, or sensationalism. It is closer to the quiet pleasure of realizing, “I didn’t know I was interested in this.” A job seeker may unexpectedly read about AI and recruitment. A young parent may discover a story about children’s media habits. A tired worker may click on an article about the four-day workweek. These are often new topics connected to latent interests: interests that were already there, but had not yet been activated.

That is why serendipity can make news consumption feel more natural and engaging, even when people were not intending to read news in the first place. It brings discovery back into a routine media environment. News is not only about crises or obligations. It can also help us notice emerging interests, name hidden concerns, and follow curiosity before it fades.

Many digital systems are good at mirroring us. They learn from yesterday’s clicks and give us more of what we have already shown interest in. This is useful, but it has limits. People are not only made of expressed preferences. We also have latent interests: topics we may care about, but have not searched for, clicked on, or clearly recognized yet. Our findings suggest that serendipitous news discovery can help bring these latent interests to the surface. When participants encountered news that was both unexpected and useful, they reported greater surprise, and that surprise was associated with stronger intentions to continue consuming news. Rather than responding only to familiar topics, they often found value in content that expanded their attention beyond their usual patterns.

Serendipitous news discovery matters because it can reveal those latent interests. Our study suggests that this process is not merely theoretical. Participants who reported stronger experiences of serendipitous discovery also reported higher levels of surprise, and this surprise was associated with a greater willingness to continue consuming news. Statistical analyses showed that surprise significantly mediated the relationship between serendipitous discovery and future news consumption intentions. In other words, serendipity did not simply make people notice an article; it helped create an emotional response that encouraged further engagement with news.

This finding is important because it identifies a specific mechanism through which serendipity operates. Previous research has often assumed that unexpected and useful information is beneficial, but has paid less attention to why it motivates people to keep exploring. Our results suggest that surprise plays a central role. When people encounter news that is both relevant and unexpected, they become more open to further discovery. Rather than reinforcing only existing interests, recommender systems may therefore be able to stimulate curiosity and sustained engagement by occasionally introducing users to topics that connect with interests they have not yet fully recognized.

Why this matters

This research contributes to serendipity studies by showing that serendipity is not only about finding unexpected and useful information. It is also about the emotional process that follows discovery. In news recommender systems, surprise helps explain why people continue consuming news after encountering something serendipitous. For scholars, this opens questions about emotion, incidental exposure, and algorithmic media. For practitioners, good recommendation should not only repeat what users already know they like. It should help people discover what they might care about next. Sometimes the news we did not plan to read is exactly the news that reminds us we are still curious.

Author bio

Zhixin Pu is an assistant professor in the Department of Communication at Beijing Normal-Hong Kong Baptist University. Dr. Zhixin Pu’s research examines how the structural design of news recommender systems and human-AI interfaces influences psychological and behavioral outcomes. Grounded in a communication perspective, her work explores AI transparency through explainable AI (XAI) to understand how system designs can be leveraged for social good, ultimately investigating how designers can intentionally structure algorithmic discovery to cultivate user curiosity and promote societal well-being.

Further reading on this topic

Binst, B., Michiels, L., & Smets, A. (2025). What Is Serendipity? An Interview Study to Conceptualize Experienced Serendipity in Recommender Systems. In Proceedings of the 33rd ACM Conference on User Modeling, Adaptation and Personalization (pp. 243–252). https://doi.org/10.1145/3699682.3728325 [also featured as Serendipity Paper]