Google: Search is changing, so here's how you'll now see news search results

 Google is using its latest machine-learning techniques to improve news suggestions for mobile users.

Google Search mobile users in the US looking for a specific news topic will now see a carousel of relevant stories at the top of the results page, plus new related coverage.

That related coverage to top stories can be seen in additional cards beneath each main card in the carousel. This feature should make it easier for users to make informed decisions about which articles to click through to, according to Google.  

"Within each story, the new structure will make more room for high-quality content – beyond just the most recent coverage – as well as more diverse sources, to bring more context and perspective to the day's news," it says in a blogpost.

Google recently revealed how its new BERT machine-learning generated models are improving how Search understands what humans are looking for when they type a query.    

BERT is short for Bidirectional Encoder Representations from Transformers, a technique Google has used to understand about 10% of search queries in English in the US. 

These same BERT models are being used as the basis for what Google sums up as its "story-understanding technology", which maps "people, places and things involved in a news story, and then draw connections between them". 

The purpose of this mapping is to create groups of news articles, and just as Google is attempting with search, to get users more quickly from question to answer. This is the feature that's rolling out for search users in the US using English today.

The new AI tech for news on mobile was announced alongside an update to the Google News Initiative, its $300m commitment over three years to support journalism in an internet dominated by Facebook and Google. Part of the plan includes Google's efforts to help publishers convert viewers into paid subscribers. 

Google on Wednesday said it was developing new machine learning-based paywall tech that may benefit publishers in the future. 

"We're applying artificial intelligence to understand how readers engage with content and ultimately make the decision to subscribe. We'll begin developing this technology in December, and intend to share learnings with the broader media industry," said Kai Ridenoure of the News Lab, Google News Initiative.

The company also has announced its other contribution to the future of integrity of information and democracy in the US via its support of a bipartisan resolution introduced this week to ensure the US 2020 Census is fair and accurate. 

The once-a-decade census is important because the data is used to decide how $800bn in annual federal spending is allocated to states and local communities in future.       

Google and YouTube's main objective is to "prevent bad actors from abusing our services to spread misinformation, or to conduct fraudulent activity around the census such as phishing or other scams".