Google has tweaked the algorithm behind 'featured snippets' to prioritize sources with up-to-date information when people ask Search questions about current and upcoming events.
Google's featured snippets are the boxed snapshots of information from the web at the top of search results that usually appear when people ask questions.
The company launched the feature in 2014 but has updated it to address quality issues after being caught returning fake news above search results and via Home smart speakers, which source answers to questions from the feature. Google detailed a few of the quirks and glitches the feature had in a blog last year.
SEE: Sensor'd enterprise: IoT, ML, and big data (ZDNet special report) | Download the report as a PDF (TechRepublic)
Despite its problems, it's still a handy feature when you need quick, short answers and Google sees it as a useful service as more searches move to mobile and voice-based queries.
The latest update to feature snippet's algorithm aims to improve Google's understanding of what information remains useful over time and what becomes stale quickly. So now when there are queries about information where freshness is important, it will scour the web for up-to-date pages.
"At the core of Search is language understanding, and our systems don't understand language the same way humans do," writes Pandu Nayak, Google Fellow and vice president of Search.
"This is why we're constantly developing new ways to better understand your searches and provide relevant results, especially in cases where there is useful context that is implied, like whether freshness matters."
Nayak provides a few examples of how the algorithm update should be helpful in surfacing more relevant information when current information matters.
For example, asking Google about upcoming school holidays previously displayed information about last year's holidays, whereas the new version should show key holiday dates for the next year.
When searching for information about an upcoming event, like a TV premier, it should provide featured snippets from sources with more specific information from pages that include facts about the date the show is available.
And for current events, such as a food recall, the old algorithm appeared to choose the most authoritative source of information, even if it was dated. The new algorithm could pick a more current news source with information about the most recent food recall.
For questions that have answers that don't change over time, such as "Why is the sky red at sunset?", the algorithm would pick an older page.