House of Lords Communications and Digital Select Committee inquiry: The future of news: impartiality, trust, and technology
Executive Summary
● The internet has resulted in more people consuming more news from more sources than ever before. Media plurality is part of Google’s design. Compared to other news sources, such as TV, radio or social networking sites, Google’s design contributes substantially to the range of views, sources and opinions offered to the user. Google users are shown multiple sources for the same story on platforms like Google Search, the “News” tab on Google Search and Google News. They see news from local, national and international sources.
● Google is one of the world’s largest financial supporters of journalism. We pay for news content from over 285 UK news titles through Google News Showcase content licensing agreements. These include the UK’s national, regional and even independent publishers. Publishers keep between 70p-95p in every £1 they make using Google Ad Manager and Google Search sends billions of free visits to news websites estimated to be worth between £0.03p-0.05p each. The Google News Initiative provided more than
$18m in grants to UK news organisations between 2018-2020. This is an important part of maintaining a healthy and pluralistic news ecosystem.
● We believe that AI is a foundational and transformative technology that will provide compelling benefits to society through its capacity to assist, complement, empower, and inspire people in almost every field of human endeavour. Google has been at the forefront of artificial intelligence (AI) research for many years and has used AI to improve products relied on by billions of people. Our teams across Google, including our AI researchers at Google DeepMind and Google Research, have achieved breakthroughs like AlphaGo and AlphaFold; pioneered advances in transformer models; and published more than nine thousand research papers, including many of the leading contributions to the field.
● Google believes AI can unlock new opportunities for news organisations. This must be done responsibly and, as such, Google is engaging closely with the news ecosystem to understand how best to balance the interests of all parties - publishers, journalists and citizens. We are prioritising approaches that will continue to send valuable traffic to news publishers.
● Publishers have told us that they want greater control over how their content is used for emerging generative AI use cases so we launched Google-Extended - a new control that all web publishers can use to manage whether they are used in our tools like Gemini and Vertex AI. We’re committed to engaging with the web, publisher and AI communities to explore additional machine-readable approaches to choice and control for web publishers, including updated web protocols.[1]
● The Google News Initiative is supporting AI training programs for journalists and research into how AI can support the news ecosystem. We have built AI tools like Pinpoint, which helps journalists and academics explore and analyse large collections of documents to help free up journalists’ capacity to focus on more impactful work.
● Google’s products are designed to surface authoritative information. Our publicly available Search Rater Guidelines set out in detail how we expect the search engine to work, with due emphasis on news. These guidelines, alongside our site “How News Works” and tools like “About This Result” provide users with substantial information about how and why we rank results the way we do. They do not take political viewpoints into account and we are not aware of any credible third party research to that effect.
● As a result of all the information we make publicly available about our product design and how we rank news content, we do not believe that Ofcom requires further statutory powers on media plurality with regard to Google’s products, based on the reasons set out in their 2022 discussion paper.
People are consuming more news from more sources than ever before
The internet has led to an unprecedented boom in the public consumption of news, from a much greater variety of sources than ever before. It has also enabled publishers to reach a much larger audience at a much lower cost. The result is that citizens now have access to more news from a much greater variety of sources.
Ofcom’s 2021 news consumption report shows this greater diversity as consumers now gain their news from an average of 8.8 news sources across all platforms (up from 6.7 in 2018).[2] The 2021 report also finds that when using the internet to consume news, the average number of news sources is 4.8, wherein the majority comprise news websites. The transformation in news consumption, marked by innovation and lower entry barriers, has shaped a diverse and evolving media landscape. According to Ofcom's 2023 report, 96% of UK adults engage with news, predominantly through broadcast TV (70%) and online sources (68%).[3]
Accessing news online has also had an impact on how people read and digest news. Instead of reading through an entire newspaper, people are now as likely to read one article at a time, at different times of the day. Whereas newspapers were traditionally read during leisure time, online news is read throughout the day. Where people once read a newspaper to get information about news, sports, business, entertainment and so on, there are now many more resources dedicated to these specific focus areas, likely contributing to the 8.8 average news sources per person in Ofcom’s 2021 report.
Despite the long-term decline in print newspapers, the internet - particularly social media - plays a significant role in news dissemination. The declining use of Google News as a means to identify and engage with news sources, from 17% in 2018 to 13% in 2023, indicates a highly competitive and diversifying environment with a range of players from established broadcasters like the BBC, to newer platforms like TikTok and WhatsApp.
While online news consumption has increased, news continues to represent a very small part of Google Search queries. In the UK, news-related queries accounted for under 2% of total queries on Google Search.
Media plurality is both explicitly and implicitly part of Google’s design
The internet, including search engines like Google Search, has significantly changed how readers access and consume news stories, and how news outlets reach readers. Google has four main surfaces in which users can consume news: Google Search (e.g. the “Top Stories” shelf in Search results), News Mode (in Search), Google News and Discover[4] (including Google News Showcase, panels from which can appear in the latter two).
Ensuring people gain access to a plurality of views is ingrained in the way we approach news. Our stated purpose in our news products is to empower everyone to understand the world through access to high quality news from a variety of perspectives. News experiences across Google are built to help users easily find news from a diversity of trusted sources so they can stay up-to-date and informed on the stories that matter to them. We believe that part of understanding the news is learning from multiple points of view and being aware of a story’s broader context. Google’s news products connect users with a wide range of sources including local community, municipal, national, and international.[5]
An example would be changes made in 2019 that allowed us to remove the requirement for news publishers to complete an application and undergo a guideline and content policy compliance review to be included in Google News products and services. Improvements to our Search Rater Guidelines and algorithms means we are better able to detect authoritative and timely news content. Consequently, publishers no longer need to proactively apply, which means we have been able to include more small and local news outlets in Google News, while continuing to maintain a high bar for quality.
Another example would be “Full Coverage”, a feature launched in Search in 2021. This added the ability for users to get greater depth and breadth of coverage through tapping into a Full Coverage page after scrolling to the end of the top stories carousel or by selecting “More news on...” right below the carousel.[6]
Media plurality is, therefore, an explicit part of the design principles of Google’s news products.
Compared to other news sources, such as TV, radio or social networking sites, Google users are shown multiple sources for the same story on platforms like Google Search, the “News” tab on Google Search and Google News. This is an important difference in the context of media plurality. On Google, multiple news sources and viewpoints are incorporated into a users’ immediate moment of consumption, providing a range of perspectives that the individual can choose to engage with. As such, media plurality is implicitly part of Google’s design and contributes substantially to the range of views, sources and opinions offered to the user.
Google also seeks to optimise for local news, including a number of product developments over the years. These include:
● Greater prominence for local news in Google News: This includes a local news panel and the ability for users to select specific locales about which they would like to see news.[7]
● Local news Showcase panels in Search: This feature is only available for local news publishers and provides special prominence for local news publishers Showcase panels in Google News .[8]
● A new local news carousel in Search: Announced in November 2021, a dedicated carousel of local news stories is triggered when Google finds local news coverage relevant to your query. The feature helps local publishers by adding another way for users to encounter their reporting.[9]
● A new label for highly cited sources: Building on the 2019 update below, this will help elevate local news reporting which can often be overtaken by national news outlets. By flagging the citations, it draws attention to the original reporting.[10]
● A 2019 algorithm update to better capture original reporting: Many important news stories are often broken by local news outlets first, before being reported by national titles. This change was aimed at ensuring local news stories get due prominence that recognises the valuable work they do in breaking new stories.[11]
Google supports the UK news environment and media plurality by helping publishers’ financial sustainability
Google works with publishers of all sizes to help them get found online, and to build tools to help even the smallest publisher succeed. With regards UK news publishers, Google does this by:
● Paying for news content in Google News Showcase: We currently licence content from over 280 of the UK’s leading regional and local news publications from publishers including Daily Mail, News UK, The Telegraph, The Guardian, REACH, NationalWorld, Newsquest, Iliffe Media, Archant, DC Thomson and the ICNN. These cover all parts of the UK.
● Driving millions of free clicks to local news sites by optimising for local news. News sites choose whether Google is able to provide its users with links to their content. They do so to expand their readership and reach people that may not otherwise have engaged with their content. All web referral traffic is estimated to be worth approximately £500m a year to UK publishers.[12] By contrast, in 2020, Google generated <$20 million in revenue (not profit) from clicks on ads against possible news-related queries in the UK.
● Providing advertising tools to help local news publishers monetise their content. News publishers keep between 70-95% of the revenues generated through these tools.[13]
● Enabling technology and business model innovation in local news through the Google News Initiative (GNI): To date, the GNI has provided over $18 million to over 80 projects in the UK through partnerships with news organisations. This includes ongoing projects with the News Media Association, including the Digital Revenue Labs, which helps local publishers unlock greater digital revenues, and the innovative Public Notices Portal, a digital platform to maintain important local publisher revenues from public notices.[14]
This combination is designed to help publishers develop sustainable business models in the digital age. From increased revenue through our search engine and advertising tools to financial support for bespoke consultancy services, Google aims to provide news publishers with the security and skills to adapt their business models and grow their revenues. As the former CEO of Archant noted on The Media Show, support from Google gave Archant “space to breathe” by de-risking their business transformation.[15] Our focus on transitioning to sustainable businesses reflects Google’s belief that this is the best way to ensure the long-term viability of news, a view echoed in The Cairncross Review.
There are signs of resilience and financial benefits arising from how publications harness digital platforms. Despite challenges in audience engagement, circulation and trust the 2023-24 WAN-IFRA World Press Trends Outlook report found publishers more upbeat about year-on-year revenue growth, expected at of 18.5% in 2024, up from a predicted 15.2% in 2023, 16.4% in 2022 and 7.3% in 2021.[16] Recent years have also seen resilience among UK publishers with The Times announcing record profits in March 2022 and editor, John Witherow, saying it is a “golden era for journalism”.[17] These are welcome signs of success and show the widespread value in online advertising to the sustainability of news publishers in the UK.
By funding access to publishers’ content, online advertising is critical to the enduring success of a pluralistic UK media and healthy information environment.
As shared with the Committee in its recent inquiry into LLMs, Google believes AI can amplify and augment human creativity, unlocking new opportunities for news organisations. This must be done responsibly and, as such, Google is engaging closely with the news ecosystem to understand how best to balance the interests of all parties - publishers, journalists and citizens.
We are prioritising approaches that will allow us to continue to send valuable traffic to web publishers, including news publishers. We’ve heard from web publishers that they want greater choice and control over how their content is used for emerging generative AI use cases. That is why we announced Google-Extended, a new control that web publishers can use to manage whether their sites help improve Gemini (previously Bard) and our Vertex AI generative APIs, including future generations of models that power those products.[18] And it’s why we’re committed to engaging with the web and AI communities to explore additional machine-readable approaches to choice and control for web publishers.[19]
Through the Google News Initiative, we are supporting training programs for journalists — so they can use AI in their work — and research into how AI can support the news ecosystem.[20] In addition, we have built research tools like Pinpoint, which helps journalists and academics explore and analyse large collections of documents.[21] Recently, a study by JournalismAI, a global initiative run through LSE’s journalism think tank Polis with the support for our Google News Initiative, showed that almost three quarters (73%) of news organisations surveyed believe generative AI applications, such as Gemini (previously Bard) or ChatGPT, present new opportunities for journalism. Some respondents noted that AI can free up journalists’ capacity for more creative work by taking on time-intensive tasks such as interview transcription and fact-checking.[22]
These findings are echoed in the 2023-24 WAN-IFRA World Press Trends Outlook report. It found that:
● AI was the biggest priority technology/product for investment, with 87% saying it was a key area for investment over the next 12 months, followed by data analytics and intelligence (86%), video (79%) and audio/podcasts (74%).
● Just over a third (34%) were very optimistic that generative AI presented opportunities for their business, while 58% were somewhat optimistic and 8% were not optimistic at all.
● At the time of the survey 67% felt their business was poorly prepared to take advantage of those opportunities, with 16% each saying they were either not prepared at all or well prepared.[23]
Google remains committed to working closely with the news ecosystem to maximise the potential from emerging AI technologies. Google’s broader approach to developing AI is governed by three ambitions - to be bold, to be responsible and to develop technologies together. We will continue to work closely with the news ecosystem to meet these goals.
Evaluation - providing access to authoritative information
Many of the questions in the call for evidence relating to trust and impartiality are adjacent to Google’s role in the ecosystem. That said, we wanted to set out how we ensure users have access to authoritative content, our approach to ranking and the lack of evidence of political bias in Google Search.
We place huge care in showing users trusted and authoritative results. The quality of these results is reflected in the fact that search engines are regularly cited as the most trusted media source, e.g. in consecutive years of the Edelman Trust Barometer survey. Similar findings are echoed in research by the Reuters Institute, including that people often use search engines like Google to verify or investigate information they find on other platforms, helping them to develop their own abilities in understanding what is trustworthy or not.[24]
Ranking: an overview
Google’s algorithms consider the following factors when ranking news in order to deliver authoritative results for users:
● Relevance: Relevance to a search term is a key factor in determining what a user sees for query-based experiences like “Top stories” in Google Search. A news article is relevant if it has the information a user is looking for. The most basic signal is when an article contains the same keywords as a user’s search.
● Prominence: Prominence is a way to identify noteworthy news stories. For example, Google’s News algorithms take into account if publishers are writing a lot of articles about a particular news story and are featuring that coverage prominently on their sites, as well as how much a story or article is trending.
● Authoritativeness: Authoritativeness signals help prioritise high-quality information from the most reliable sources available. To do this, Google’s systems are designed to identify signals that can help determine which pages demonstrate expertise, authoritativeness and trustworthiness on a given topic. Those signals can include whether other people value the source for similar queries or whether other prominent websites on the subject link to the story.
● Freshness: Freshness refers to how recently the article was published and how important to this story having the freshest content is. When news is happening, Google’s algorithms may determine that an article with up-to-date information is likely more useful than an older one.
● Location: Location may influence which article a user sees. Google uses a user’s location to help find relevant content, such as local news stories in Google News.
● Usability: Usability assesses how easy it is to view content on a site, such as whether the site appears correctly in different browsers; whether it is designed for all device types and sizes, including desktops, tablets, and smartphones; and whether the page loading times work well for users with slow Internet connections.
More information is available on our site How News Works.
Search Rater Guidelines provide detailed transparency on Google’s approach to ranking news
At the heart of Google’s commitment to ranking transparency lies our publicly available Search Rater Guidelines.[25] This is a 170 page public document that sets out, in detail, how we expect our search engine to perform. In short, these guidelines are a clear description of what we value in content when ranking. Each year, we conduct 100,000s of experiments which are verified by 10,000 independent search raters around the world. They are asked to test our results against how we expect them to be based on these guidelines. A shortened version of the guidelines is also available to ensure ease of access for the public.[26]
At the heart of Page Quality rating is Expertise, Authoritativeness, Trustworthiness (E-A-T). When it comes to news, high E-A-T news articles should be produced with journalistic professionalism—they should contain factually accurate content presented in a way that helps users achieve a better understanding of events. High E-A-T news sources typically have published established editorial policies and robust review processes.
To ascertain whether a website contains high quality news pages, raters are asked to consider how the publication features locally, alongside issues like whether the newspaper has a positive reputation for its reporting, its history in winning awards, the relationship between the main content of a news page and advertising e.g. are ads clearly labelled, is an op-ed labelled as an op-ed.
The Guidelines clearly instruct raters to use the highest rating, “very high quality,” for original news reporting “that provides information that would not otherwise have been known had the article not revealed it”. It also notes “very high quality news content will (often) include a description of primary sources and other original reporting referenced during the content creation process”.
In addition to recognising individual instances of original reporting at the page level, we also ask raters to consider the publisher’s overall reputation for original reporting. The relevant section (2.6.1) instructs raters to consider the overall reputation of the publication such as whether it has won journalistic awards or a history of high quality original reporting.
Conversely, the Rater Quality Guidelines also offer guidance on how obscured content, deceptive purposes, and auto-generated or copied content may be suggestive of lower quality.
When it comes to automatically generated content, our guidance has been consistent for years. Using automation—including AI—to generate content with the primary purpose of manipulating ranking in search results is a violation of our spam policies.
Google has many years of experience dealing with automation being used in an attempt to game search results. Our spam-fighting efforts—including our SpamBrain system—will continue, however spam is produced.
This said, it's important to recognize that not all use of automation, including AI generation, is spam. Automation has long been used to generate helpful content, such as sports scores, weather forecasts, and transcripts. AI has the ability to power new levels of expression and creativity, and to serve as a critical tool to help people create great content for the web.
As explained above, those seeking success in Google Search should be looking to produce original, high-quality, people-first content demonstrating E-A-T. This helps to maintain high quality results for users. More detail on AI-generated content is set out here.
A shortened version of the guidelines is also available to ensure ease of access for the public.[27]
Google’s algorithms do not take political viewpoint into account
We strongly reject any claims that Google Search and News are politically biassed. Our systems do not take the political viewpoint of a webpage into account when ranking - they are simply not built to do that. As mentioned, the Search Rater Guidelines provide a substantial degree of transparency into how we rank content, along with the material available on our Webmaster Guidelines.
The Search Rater Guidelines are designed to prevent bias entering into the testing process. They explicitly state that “[r]atings should not be based on your personal opinions, preferences, religious beliefs, or political views''. When we make improvements to our systems, we look for ways to improve Search not just across a few queries, but for a broad range of topics and queries. These changes do not target specific sites and, as we have said many times, our products, processes, policies do not take political point of view into account.
We are not aware of any credible third party study that finds evidence of political bias on Google News or Google Search. In fact, a number of studies find the opposite. For example:
● Two studies by The Economist evaluated claims of bias in Google News results. They found no evidence of ideological bias. They draw a clear conclusion: “Google rewards reputable reporting, not left wing politics”.[28]
● An extensive study by academics at Stanford University draws a similar conclusion, examining over 4 million sites for evidence of political bias in search results.[29]
A detailed list of other third party studies is included in the appendix.
YouTube
How YouTube surfaces news
Alongside searches, YouTube may surface news content in other ways. These include:
● Top News shelf in Search: A Top News shelf in Search shows up if you look for news and we have relevant content from news sources that follow YouTube’s Community Guidelines and other YouTube policies that may apply. We also take into account other signals such as channel quality and coverage of recent and relevant news stories.
● Top News shelf on a users’ homepage: A Top News shelf on a users’ homepage may be shown if they watch or search for news content. The Top News shelf will show up regardless of age, and will include relevant content from news sources that follow YouTube’s Community Guidelines and other YouTube policies that may apply. We also take into account other signals such as channel quality and coverage of recent and relevant news stories.
● Breaking News shelf on the homepage: A Breaking News shelf appears on the YouTube homepage, except for users under 18. When there is a significant news event, the Breaking News shelf on the homepage can be found by scrolling down the YouTube homepage.
YouTube supports news organisations to grow their audiences and revenues by providing a range of trainings, resources and also to foster informed viewers.[30] Features such as “Developing News” give users important context on breaking news stories. YouTube also engages a range of fact checkers and includes fact checking panels to promote authoritative and trustworthy content.[31]
YouTube has explicit guidelines prohibiting misinformation
YouTube does not allow misleading or deceptive content that poses a risk of egregious harm. When it comes to misinformation, YouTube seeks robust facts on which to base its policies. For example, for COVID-19 medical misinformation policies, we rely on expert consensus from both international health organisations and local health authorities, such as the NHS.
Misinformation policies include:
● Suppression of census participation: Content aiming to mislead census participants about the time, place, means, or eligibility requirements of the census, or false claims that could materially discourage census participation.
● Manipulated content: Content that has been technically manipulated or doctored in a way that misleads users (beyond clips taken out of context) and may pose a serious risk of egregious harm.
● Misattributed content: Content that may pose a serious risk of egregious harm by falsely claiming that old footage from a past event is from a current event.
● Promoting dangerous remedies, cures, or substances: Content that promotes harmful substances, treatments, or substances that present an inherent risk of severe bodily harm or death.
● Contradicting expert consensus on certain safe medical practices: Content that contradicts local health authorities’ or WHO guidance on certain safe medical practices.
The policies are developed in partnership with a wide range of external experts and are enforced using a combination of content reviewers and machine learning to remove content that violates the policies as quickly as possible.
YouTube’s “Four R’s” approach
YouTube has a “four R’s” approach to information quality:
● Raise: We promote content from trusted sources, like major media outlets, in search results and information panels. In 2017, we started to prioritise authoritative voices, including news sources like major UK news outlets, for news and information queries in search results and “watch next” panels. For searches on newsworthy topics, on average, 93% of the videos in global top 10 results come from high-authority channels.
● Remove: We actively take down content which violates our policies through a combination of algorithmic and human moderation. In our latest quarterly transparency report[32] (Q2 2022) data shows that of the 4.5 million videos YouTube removed for violating our Community Guidelines last quarter, 93% were first detected by our automated flagging systems. We also set out in the same report how many of the videos removed were done so for the reason of misinformation (in Q2 2022, it was 122,660 videos or just under 3% of the overall violative videos over the period).
● Reduce: For content which comes close to violating policies but does not quite cross the line, we work to reduce its discoverability on YouTube to limit views. Globally, consumption of borderline content or harmful misinformation videos that come from our recommendations is significantly below 1% of all consumption of content from recommendations.
● Reward: News organisations can generate revenue from their YouTube content through advertising, as part of the YouTube Partner Program (YPP). They must follow our monetisation policies, including Google AdSense program policies and YouTube’s Advertiser-Friendly Guidelines. Many of the UK’s leading publishers are part of YPP.
YouTube has also launched a number of approaches to drive consumption of authoritative news content, for example:
● Providing reliable information faster for breaking news: Reliable information becomes especially critical as news is breaking and as events are unfolding, it can take time to produce high-quality videos containing verified facts. Consequently, YouTube started providing short previews of text-based news articles in search results, along with a reminder that breaking and developing news can rapidly change. We’ve also introduced Top News and Breaking News sections to highlight quality journalism. This year alone, we’ve seen that consumption on authoritative news partners’ channels has grown by 60 percent.
● Providing context to users: Sometimes a video alone does not provide enough context to viewers about what they are watching. We want to make sure that people who watch videos about topics prone to misinformation are provided additional information while viewing. To that end, we’ve designed a variety of information panels that target different types of context, such as general topics and recent news prone to misinformation, or about publishers themselves. For example, when people watch videos that encourage viewers to skip the MMR vaccine, we show information panels to provide more basic scientific context, linking to third-party sources. Or if people are viewing news videos uploaded by a public broadcaster or a government-funded news outlet, we show informational notices underneath the video about the news outlet. Collectively, we’ve delivered more than 3.5 billion impressions across all of these information panels since June 2018.
How well is regulatory oversight working? Are any changes needed, for example: a) In the way Ofcom oversees due impartiality and the extent of its remit? b) In the way Ofcom oversees media plurality?
Ofcom’s discussion paper into media provides the following motivation for exploring further statutory powers around media plurality.
As intermediaries increasingly play the role of gatekeepers, curating or recommending news content to online audiences, it is not clear that people are aware of the choices being made on their behalf, or their impact. To better understand the implications of these changes…we set out to examine the possible impacts of the growth of online news, and the role of online intermediaries in particular, on media plurality, and what, if any, regulatory changes may be necessary to maintain and secure it.[33]
As set out above, Google’s provides significant transparency to users why they are seeing what they see. From the detailed Search Rater Guidelines to tools like “About This Result”, we work hard to provide people with meaningful information about what they are seeing.
As such, we do not believe further Ofcom powers are necessary with regards to Google’s products.
Conclusion
Google is absolutely committed to playing its part in supporting a healthy news ecosystem. We support media plurality and access to trusted information explicitly in our product design. As one of the world’s largest financial supporters of journalism, we support media plurality through contributing to financial sustainability. Our products are designed to surface authoritative information, with due emphasis on news, as set out in the Search Rater Guidelines. And we make substantial information about how and why we rank results the way we do readily available to users. We continue to be one of the world’s foremost AI companies and believe these technologies offer opportunities for the news ecosystem. Critically, they can also deliver valuable societal benefits and, with the right policy conditions, allow the UK to be a world leader in harnessing AI for growth.
We look forward to further engaging with the Committee.
Appendix - third party research on Google, media plurality and trust in news
Misinformation
Leibniz Institute for the Social Sciences: Where the Earth is flat and 9/11 is an inside job: A comparative algorithm audit of conspiratorial information in web search result
● The Institute conducted a comparative algorithm audit of presence of conspiratorial information in top search results across five search engines such as Google, DuckDuckGo, Yahoo, Bing and Yandex. Their research found that “all search engines except Google consistently displayed conspiracy-promoting results and returned links to conspiracy-dedicated websites in their top results, although the share of such content varied across queries.”[34]
City University of Hong Kong: Can “Googling” correct misbelief? Cognitive and affective consequences of online search
● Researchers at the City University of Hong Kong conducted two experiments on misinformation about ethnic minorities in Japan. Control groups were asked to search widely shared, prejudiced misinformation about an ethnic minority in Japan. Both studies found that:
Sussex Medical School: Online information of vaccines: information quality, not only privacy, is an ethical responsibility of search engines
● A study by researchers at Sussex Medical school was conducted to investigate the relationship between search engines’ approach to privacy and the scientific quality of the information they return, given the role search engines can play in spreading low-quality health information, such as anti-vaccine websites. The study analysed the first 30 webpages searching the phrase “vaccines autism” in English, Spanish, Italian, and French. The results found that Google is consistently returning less misinformation than “alternative” search engines (Duckduckgo, Ecosia, Qwant, Swisscows, and Mojeek) but also other commercial engines (Bing, Yahoo) often return more anti-vaccine pages (10–53%) than Google.com (0%).[36]
Polarisation and algorithmic bias
Reuters Institute/University of Oxford: Echo Chambers, Filter Bubbles, and Polarisation: a Literature Review
● The literature review examines social scientific evidence regarding the “existence, causes, and effect of online echo chambers” in the context of concerns about digital platforms, contributing to polarisation in societies generally and, in relation to scientific topics, specifically. The research found that “automated serendipity and incidental exposure mean that relying on search engines, social media, and other digital platforms using algorithmic ranking leads people to slightly more diverse news - the opposite of what the filter bubble hypothesis posits.”[37]
Stanford University: Search results don’t have a political bias
● In response to questions about platforms’ influence over what news and information people see online, and the potential for political bias and censorship in search engine results, researchers reviewed “the first page of Google search results for every candidate running for federal office in the 2018 US election over a six-month period. After a systematic audit of about 4 million URLs scraped from the search engine, they found that the results don’t exclude sources from either end of the political spectrum. For the most part, the researchers found that the news sources most commonly held a relatively centrist point of view.”[38]
The Economist: Google rewards reputable reporting, not left-wing politics
● This study sought to understand whether there was any evidence of ideological bias in Google’s news tab. It concluded that “Google rewards reputable reporting, not left-wing politics”.
Michigan State University, University of Ottawa, University of Oxford: Search and Politics: The Uses and Impacts of Search in Britain, France, Germany, Italy, Poland, Spain, and the United States
● This study sought to address issues over whether biases embedded in the algorithms that drive search engines and social media have major political consequences, such as creating filter bubbles or echo chambers. In the study they asked internet users how they use search, social media, and other important media to get information about political candidates, issues, and politics generally, as well as what difference it makes for individuals participating in democratic processes - the survey was conducted in Britain, France, Germany, Italy, Poland, Spain, and the US. The findings cast doubt on the filter bubble argument - instead finding that it is overstated, as “internet users expose themselves to a variety of opinions and viewpoints online and through a diversity of media. Search needs to be viewed in a context of multiple media.”[39]
Northeastern University: Auditing Partisan Audience Bias within Google Search
● To assess whether there is any truth to the notion that online platforms have a systemic influence on the democratic process, researchers used mixed-methods algorithm audit of partisan audience bias and personalisation within Google Search. Following the inauguration of Donald Trump, researchers asked 187 participants “to complete a survey and install a browser extension that enabled [them] to collect Search Engine Results Pages (SERPs) from their computers. To quantify partisan audience bias, [they] developed a domain-level score by leveraging the sharing propensities of registered voters on a large Twitter panel.” In the paper, they reported that they found little evidence for the “filter bubble” hypothesis. Instead, they found that the “direction and magnitude of overall lean varied by search query, component type (e.g. “answer boxes”), and other factors.”[40]
Hamburg University of Applied Sciences: A Comparative of Source Distribution and Result Overlap in Web Search Engines
● This paper reports the findings of a study examining the difference between results retrieved by four major web search engines. Researchers compared the top 10 results from Google, Bing, DuckDuckGo, and Metager, using 3,537 queries generated from Google Trends from Germany and the US. The findings of the study showed that “Google displays more unique domains in the top results than its competitors, and Wikipedia and news websites are the most popular sources overall.”[41]
Institute for Strategic Dialogue: 404: Reliable Information Not Found
● This research evaluates the performance of three social media platforms in surfacing reliable news about the war in Ukraine. Using a methodology of searching the keyword ‘Ukraine’ and analysing the top results, ISD found that pro-Kremlin accounts, along with unverified sources, were highly ranked on some major platforms. Only YouTube, however, was found to raise media sources and verified profiles. Specifically, it found that “YouTube’s results comprised almost exclusively media sources (38/40), and contained zero pro-Kremlin accounts”.[42]
February 2024
20
[1] See Danielle Romain, A principled approach to evolving choice and control for web content, The Keyword (Jul. 6, 2023), https://blog.google/technology/ai/ai-web-publisher-controls-sign-up/
[2] Ofcom, News consumption in the UK, 2021
[3] https://www.ofcom.org.uk/data/assets/pdf_file/0024/264651/news-consumption-2023.pdf p.7
[4] Discover is a part of Google Search that shows people content related to their interests, based on their Web and App Activity. See here for more detail.
[5] See Google How News Works for more detail.
[6] See Get the full news story with Full Coverage in Search for more detail.
[7] The Keyword, Read all about it: A new look for Google News, 22 Jun 2022
[8] The Keyword, Dive deeper into local news with News Showcase, 9 Feb 2022
[9] The Keyword, New tools and features to support local news, 16 Nov 2021
[10] The Keyword, A new label for highly cited sources, 31 Mar 2022. See also Hold the Front Page, 08 Sep 2022, Editorial boss welcomes Google move to boost original journalism
[11] The Keyword, Elevating original reporting in Search, 12 Sep 2019
[12] Deloitte (2019) The impact of web traffic on revenues of traditional newspaper publishers
[13] Google (2020) A look at how news publishers make money with Ad Manager
[14] See e.g. News Media Association (2023) Local News Sector Announces Full Launch Of Public Notice Portal
[15] BBC, The Media Show, 1 September 2021
[16] See World Association of News Publishers, World Press Trends Outlook 2022-2023 (Sept. 3 2023) https://wan-ifra.org/insight/report-world-press-trends-outlook-2022-2023/
[17] The Times (3/11/2021) A ‘golden age’ for serious journalism
[18] See Danielle Romain, An update on web publisher controls, The Keyword (Sept. 28, 2023), https://blog.google/technology/ai/an-update-on-web-publisher-controls/
[19] See Danielle Romain, A principled approach to evolving choice and control for web content, The Keyword (Jul. 6, 2023), https://blog.google/technology/ai/ai-web-publisher-controls-sign-up/
[20] See The London School of Economics and Political Science, JournalismAI Home Page, Dept. of Media and Communications, https://www.lse.ac.uk/media-and-communications/polis/JournalismAI (last visited Oct. 27, 2023).
[21] See Brendan McCarthy, How technology powered a Pulitzer Prize-winning investigation, The Keyword (Nov. 16, 2021), https://blog.google/products/news/boston-globe-pinpoint-pulitzer-prize/
[22] See Charlie Beckett, How AI is generating change in newsrooms worldwide, The Keyword (Sept. 20, 2023)
https://blog.google/technology/ai/how-ai-is-generating-change-in-newsrooms-worldwide/
[23] WAN-IFRA (2024) World Press Trends Outlook 2023-24
[24] See e.g. Edelman Trust Barometer 2021, 2022, 2023 and Reuters Institute (2022) The Trust Gap.
[25] See Google Search Rater Guidelines for more detail.
[26] See Search Quality Rater Guidelines: An Overview for more
[27] See Search Quality Rater Guidelines: An Overview for more
[28] The Economist (2019) Google rewards reputable reporting, not left-wing politics.
[29] Metaxa D, Sung Park J, Landay J, Hancock J (2019)N Search Media and Elections: A Longitudinal Investigation of Political Search Results, proceedings of the ACM on Human-Computer Interaction, Volume 3 Issue CSCW. Article No.: 129.
[30] See e.g. this case study where YouTube helped The Economist drive a wide range of engagement including: over 1M subscribers to the channel and significantly increasing female viewership of The Economist's content (21% up from 5%).
[31] See https://news.youtube/foster-informed-viewers/ for more information.
[32] See YouTube’s Transparency Report
[33] Ofcom (2022) Media plurality and online news discussion document, p.1
[34] Urmana A, Makhortykhb M, Ulloac R, Kulshresthad J (2021) Where the Earth is flat and 9/11 is an inside job: A comparative algorithm audit of conspiratorial information in web search results: Leibniz Institute for the Social Socials.
[35] Kobayashi T, Taka F, Suzuki T (2021) Can “Googling” correct misbelief? Cognitive and affective consequences of online search:City University of Hong Kong.
[36] Ghezzi P,. Bannister P, Casino G, Catalani A, Goldman M, Morley J, Neunez M, Prados-Bo A, Smeesters P, Taddeo M, Vanzolini T, Floridi L (2020) Online information of vaccines: information quality, not only privacy, is an ethical responsibility of search engines: Sussex Medical School:
[37] Arguedas A, Robertson C, Fletcher R, and Nielsen R (2022): Echo Chambers, Filter Bubbles, and Polarisation: a Literature Review: Reuters Institute/Oxford University.
[38] Futurity (2019) Search Results Don’t have a Political Bias.
[39] Dutton H, Reisdorf B, Dubois E, Blank, G Ahmad S, Robertson C (2017) Search and Politics: The Uses and Impacts of Search in Britain, France, Germany, Italy, Poland, Spain, and the United States: Michigan State University, University of Ottawa, University of Oxford.
[40] Robertson R, Jiang S, Joseph K, Friedland L, Lazer D, Wilson C (2018), Auditing Partisan Audience Bias within Google Search: Northeastern University.
[41] Yagci N,Sünkler S, Häußler H, Lewandowski D (2022) A comparative of Source Distribution and Result Overlap in Web Search Engines: Hamburg University of Applied Sciences.
[42] ISD (2022) 404: Reliable Information Not Found