Is Google Personalized News Using AI To Create Bias?

Contents hide

1 Is Google Personalized News Biased?

2 How Google Personalized News Works

2.1 How Google Assembles News for a New Account

2.2 How Google AI Compiles News After Getting To Know You

2.3 Google’s AI Refining Your Interests From the Content You Consume

2.4 Training Google News To Show Content That You Want

3 Google News Is Just a Messenger

3.1 Google News Top Stories Media “Selection Bias”

3.2 Showing up in Google News

4 Key Takeaways

5 References:

Have you ever gone through your Google News feed and noticed almost all the news comes from the same or similar news channels? Google News seems to push similar types of news and rarely shows topics that counter the other’s narratives. Does that mean Google personalized news is creating and spreading bias amongst the public?

Google personalized news isn’t purposely using AI to create bias. There’s no evidence to suggest that Google’s AI is biased toward a specific news outlet or ideology. However, the AI appears biased as it tries to compile news that a user will like, thereby automatically aligning with the user’s bias.

This article entails a comprehensive overview of why it seems that Google personalized news is biased. To help you better understand the Google News platform, there’s also a detailed breakdown of how the platform gathers your news selection and how you can train Google News to show you the content you want.

Is Google Personalized News Biased?

Google personalized news is biased to show you news on topics you’re interested in and from sources you like. Google uses AI to determine the type of news you want and your preferred publications. It then assembles news relating to those sources and topics to increase engagement.

As a business, Google News wants users to enjoy their platform and use it more often. To do this, they’ve engineered an AI that first understands the user’s interests and then searches the internet to find the most relevant articles, documents, and posts they’re likely to engage with (click on, read, share, etc.).

So for any particular user, Google’s recommendation will be biased to their specific interests.

That said, over the past few years, there have been claims that Google News is pushing a biased narrative, where it’s more heavily distributing a certain kind of content and not showing both sides of the coin.

There have been certain instances where people with a particular political affiliation get bombarded with news and articles that promote a different agenda. But does that mean that Google’s AI is trying to sway public opinion in favor of a particular political party or ideology?

No one is sure, so to better understand why this happens from time to time, you will need to get an idea of how Google personalized news works.

How Google Personalized News Works

No one knows the exact ins and outs of how Google’s AI compiles news or any content for that matter. This secrecy is because Google’s ranking algorithm is the company’s copyrighted intellectual property, and they don’t publicly share information on how it works.

That said, Google does share an abstraction of how news and content are constructed, which we showed in our previous article “Is Google News The Same For Everyone?”

The following sections discuss how Google compiles news and other stories for its users.

How Google Assembles News for a New Account

Let’s say you create a new Google account with absolutely zero activity so that the company doesn’t know anything about your interests and likings. In this case, what type of news will Google present to you?

With access to your preferred language and region, Google will list the most relevant and popular news stories. In fact, according to Google, everyone with the same language and region will get the same news on the same subject.

Google picks the “most relevant” news by prioritizing what’s popular. If more people are reading it, Google assumes that the news is relevant and something you might also like. Second, Google picks news outlets that receive a lot of traffic and backlinks – which signals that the site is excellent and trustworthy.

So going by this information, Google is biased toward popular news coming from sites with more traffic and backlinks. Apart from this, there is no evidence to suggest other biases at play.

How Google AI Compiles News After Getting To Know You

How Google AI Compiles Your News After Getting To Know You — Google AI Machine learning algorithms compile news after reviewing your personal information.

As you engage with Google news and use Google Search or YouTube, Google’s AI will learn more about you through machine learning algorithms.

The AI will keep track of topics, sources, and locations you mark as favorites and start to initialize your personalized news feed with content that reflects your likings and interests. It will then present you with news and stories that align with what you enjoy.

However, sometimes it doesn’t work perfectly; it may make you think Google is pushing an agenda when it says it is not! Whether Google’s AI is considered “Black Box AI” is still to be determined as no one knows for sure what is in their AI algorithms. But we investigate this subject in an article titled “Is Black Box AI Dangerous?”

Google’s AI Refining Your Interests From the Content You Consume

It is important to remember that Google’s AI is working on refining your interests and likings from the content you consume. Google uses artificial intelligence machine learning to find patterns in your data.

For example, let’s say you personally like iPhones more than Android phones. You go on Google to search for the keyword – “iPhones vs. Android.” Next, you watch a YouTube video discussing “10 things people hate about iPhones.”

With access to only this data, Google’s AI may access that you don’t like iPhones and start pushing content related to Android instead. It can even double down on the “10 things people hate about iPhones” video you watched and start showing content related to the problems with iPhones.

This chain of events might give the impression that Google is anti-iPhone, pushing you towards Android – Google’s own product. Google uses different aspects of AI for understanding concepts to decipher the intentions of users. They use RankBrain, neural networks “Neural Matching,” BERT, and MUM, all analyzing different aspects of information.

But, in reality, AI is growing in its learning every day and may have interpreted your “interest” incorrectly. The problem should rectify itself as soon as the AI machine learning analyzes more data that you are reading fewer anti-iPhone articles in favor of pro-iPhone ones.

Training Google News To Show Content That You Want

If you find that Google Personalized News is showing you content you don’t like, you can adjust some settings and train the AI to tailor your personalized news to cover certain sources and specific topics you want to read.

To do this, follow the given steps for the Google News App:

Open the Google News app.
Open the personalized news section by clicking on the “For You” tab at the bottom of the screen.
If you find an article you like, click the three-dot button on its card and tap on the “more stores like this” option.
Likewise, if you find an article you don’t like, click the three-dot button and top on the “fewer stores like this” option.
If you want to stop getting news from a particular source, you can select the option – “Hide all stores from *this source*.”

This adjustment will provide Google’s AI with more direct information on what you like and dislike, improving the news curation with time.

Google News Is Just a Messenger

When you do a quick Google search, and it shows you an answer, it doesn’t mean that Google – the company, advocates for that answer.

It’s important to realize that Google isn’t a publishing company that produces and distributes information. Google News is a News Aggregator, and it is more like a library that organizes and indexes all the available publications and presents the users with information that it feels is most relevant to their query.

Google’s search algorithm is a business secret. So the real question is, how does Google determine which answer is more relevant? How does Google decide that this answer should be number one on their search results page while others go below it, and some don’t even appear on the first page?

While there are some third-party audits of the Google search algorithm for Top Stories to know whether any inherent bias exists, there is evidence to suggest that there are content-based biases.

There are studies on Google News Aggregation and its effect on news access as a whole.

Google News Top Stories Media “Selection Bias”

A study was released in 2019 by Emma Lurie and Eni Mustafaraj of the Department of Computer Science at Wellesley College. The study was called “Opening Up the Black Box: Auditing Google’s Top Stories Algorithm” and was partially funded by the National Science Foundation through grant IIS 1751087.

They ran experiments to determine what kind of news content, articles, and posts Google prefers and ranks higher.

The study indicated that “CNN’s behavior is an outlier compared to all other news publishers. That is, CNN stories have some particular features that provide them some advantage in Google’s Top stories ranking.”
The National Science Foundation Public Access Repository: “Opening Up the Black Box: Auditing Google’s Top Stories Algorithm”

There have not been many studies on the Google News Search algorithms as they are difficult to obtain, and this study has not been updated with Google Core Updates, so it is difficult to assess the current effects of news search.

Showing up in Google News

It is necessary for anyone supplying News to Google to meet Google News requirements. If you want to Show up in Google News as they describe it, Google support has Publisher Center Help with multiple links and resources.

Google also has the Publisher Center Tool that you can share your content with Google News by submitting RSS or (Really Simple Syndication) feeds, website URLs, and videos. It does not mean they will use your information, but it certainly gives you a better opportunity.

As Google says, you don’t need to sign up through the Publisher Center to be considered for Google News rankings, but it does provide certain benefits:

Monetization Opportunity
Placement Eligibility
Content and Branding Control

To be considered by Google, other factors have a lot to do with the subject matter of the content. They also look at other technical facets like readability, loading speeds, formating with proper headlines, on-page time, bounce rate, information about the domain, and backlinks, among others.

Key Takeaways

Google personalized news and its employees are likely not purposely using AI to create bias, but it is using AI to show you news biased towards your specific interests and likings. AI at this time is inherently biased from training data.

Google News aims to keep people using its platform because they are a business. They employ Artificial Intelligence learning to understand what users like and then present content aligned with their liking.

There has been a history of “gaming” the Google algorithm by optimizing a piece of content to rank for an unrelated keyword – a practice known as Google bombing. So something as big as Google is far from infallible.

However, there’s no evidence that Google, as a company, is purposely incorporating bias in its personalized news to sway the public towards a particular belief system or ideology.