Google using MUM to improve featured snippets

Posted by Edith MacLeod on 30 Aug, 2022
View comments Search News
MUM can identify consensus and false premise for featured snippets.

Google MUM in snippets.

Image: Unsplash

Google has announced a series of improvements to Search, aimed at improving the quality of search results, as well as information literacy across the web.

Google says its MUM (Multitask Unified Model) AI can now understand consensus - that is, when multiple high-quality sources on the web agree on the same fact.

“Our systems can check snippet callouts (the word or words called out above the featured snippet in a larger font) against other high-quality sources on the web, to see if there’s a general consensus for that callout, even if sources use different words or concepts to describe the same thing.”

Here’s Google’s example, showing the callout in larger font above the snippet.

Featured snippet callout.

Google describes this as a significant improvement to snippets, saying the consensus-based technique has “meaningfully improved” the quality and usefulness of featured snippet callouts.

At the same time, the AI can also detect false premise and identify where a snippet would not be useful to display.

“This is particularly helpful for questions where there is no answer: for example, a recent search for “when did snoopy assassinate Abraham Lincoln” provided a snippet highlighting an accurate date and information about Lincoln’s assassination, but this clearly isn’t the most helpful way to display this result.”

Google says their systems are now better at detecting these kinds of false premise queries. This update has resulted in a 40% reduction in the triggering of featured snippets in these cases.

In their drive to improve information literacy features on Search, Google has also rolled out updates to About this result and content advisories for information gaps.

About this result

This feature allows people to see more about a result before clicking on it, and Google says it has been used over 2.4 billion times since it was launched in 2021.

Google is now adding more context to About this result including:

  • How widely a source is circulated
  • Online reviews about a source or company
  • Whether a company is owned by another entity
  • When Google’s systems can’t find much information about a source

The feature has been added to the Google app, and is being expanded to cover eight more languages: Portuguese (PT), French (FR), Italian (IT), German (DE), Dutch (NL), Spanish (ES), Japanese (JP) and Indonesian (ID) later this year.

About this result.

Content advisories

Content advisories were introduced last year by Google for situations where a topic is evolving rapidly, or there is a lack of reliable information online about a particular subject.

It is now expanding these to cover searches where their systems don’t have high confidence in the overall quality of the results available for the search.

Google says this doesn’t mean no helpful information is available, or that a particular result is low-quality. The notices provide context about the whole set of results on the page. The searcher will always be able to see the results for their query, even when the advisory is present.

Content advisories.


For further reading, see our blog post on Google's LaMDA and MUM AI technologies.

Recent articles

Google's March 2024 updates - the impact so far
Posted by Edith MacLeod on 19 March 2024
Google is rolling out the March 2024 core and spam updates
Posted by Edith MacLeod on 15 March 2024
Interaction to Next Paint comes to Core Web Vitals
Posted by Edith MacLeod on 12 March 2024
Google says AI-generated images in Merchant Center must be labeled as such
Posted by Edith MacLeod on 4 March 2024
Microsoft’s guide to using prompt engineering for your ad copywriting
Posted by Edith MacLeod on 2 March 2024