The tendency to seek information in a way that confirms a user's pre-existing beliefs. For example, searching for "health benefits of drinking coffee" will yield search results that highlight the benefits of drinking coffee, potentially reinforcing the perception that it is a healthy behavior, a perception the user already had when they were drafting the query. Search, because it is locating resources that best match the tone and frame of a user's query, is the primary driver of this tendency online. However, confirmation bias also occurs in the way that we interact with information in any context. Our minds have a slippery tendency to forget information that doesn't align with our beliefs, and retain information that does comport with our beliefs with higher fidelity. Because this is an inherent property of our psychology, while platforms can take some steps to try to limit their contribution to this tendency, they are unable to fully neutralize it.
However, confirmation bias is a much easier problem to combat that other perception skewing effects (like filter bubbles and echo chambers). This is because a Search is a user is undertaking a self-directed query to find something out. This moment of curiosity is exactly the time where the user may be maximally interested in an alternative perspective, so search is uniquely well positioned to both cause confirmation bias, but also to combat it.