Japan
Japan’s media landscape is characterised by significant print circulation for national and regional broadsheet newspapers and a dense network of television broadcasters, including public broadcaster NHK. However, that dominance is increasingly challenged by social networks and video platforms in terms of influence and profitability. Legacy outlets have struggled to maintain their position as the primary source of information for voters in recent major elections.
The circulation of Japan’s daily newspapers continues to decrease steadily; in October 2025, total circulation fell to 24m copies, representing a 6.57% drop from the previous year. The readership is ageing and the industry's influence is waning.
At the same time, the success of legacy media’s digital transformation has been limited. Nikkei, or the Japan Economic Daily, which owns the Financial Times, achieved 1.21m digital subscribers (including affiliated media, e.g. Nikkei Prime), nearly matching its print subscriber base of 1.25m. In contrast, the digital subscriber count for Asahi Shimbun, one of the earliest adopters of digital platforms, has remained around 300,000 in recent years, despite the paper losing millions of print readers over the last decade. Meanwhile, Yomiuri Shimbun, with a circulation of nearly 6m, continues with its bundle-only policy to maintain healthy print circulation figures. Smaller regional newspapers are further behind, with 74 out of 79 local and national newspapers generating less than 5% of their revenue from digital sales (Japan Newspaper Publishers and Editors 2026). In terms of digital-born brands, Yahoo! retains a strong weekly news reach, having been a portal giant in the early days of the internet in Japan, while LINE News leverages its position as Japan’s dominant messaging app. Both offer aggregated news alongside original content.
As the challenges for newspapers continue, Asahi Shimbun’s CEO Katsu Tsunoda has clarified its stance of ‘pivoting entirely toward AI integration’.1 This involves, for example, science reporters using AI to summarise research papers and implementing systems where reporters are only alerted to high-priority late-night police press releases. However, he emphasises that obtaining primary information through reporting remains a task exclusively for humans. News app SmartNews has partnered with the YouTuber management firm UUUM to convert hobby-related videos into text for distribution using AI.2 Outside of AI initiatives, the Mainichi Shimbun, Japan’s oldest daily, has produced documentaries focusing on the national elections in 2025 and 2026.
The broadcast sector is also struggling to maintain its viewership, particularly among younger audiences. The time people spend watching live weekday TV has been declining since it was overtaken by online in 2021, according to the government’s annual Survey on Information and Communication Media Usage Time and Information. The latest figures from 2024 show that only over-60s spend more time watching live television than using the internet.3 Against this backdrop, Japan’s five major commercial broadcasters are working to expand access to their content through TVer, a free streaming service they jointly operate. The platform offers ad-supported streaming of drama and entertainment programmes, and livestreamed terrestrial broadcasts. It currently hosts around 800 programmes and attracts 44m monthly users.
The 2025 revision of the Broadcasting Act redefined online services as a mandatory function for public broadcaster NHK. Regular users of the newly launched NHK ONE online and app are charged if they do not already pay the television licence fee. However, it remains unclear whether the new service will compensate for declining income from viewing fees.
Changes in media consumption are becoming evident in political discourse, as was the case in the 2026 snap election. The country’s first female prime minister, Sanae Takaichi, secured a landslide victory as skilfully crafted social media content portrayed her as relatable and energetic. The ruling party’s campaign video also featured her heavily and was viewed 130m times on YouTube.4 However, a Nikkei investigation found that, among all election-related videos, videos from anonymous creators accounted for 55% of the total views,5 highlighting the difficulty of detecting unverified information and foreign influence operations.
The surge of short-form video during the 2025 Upper House election and the 2026 general election presented new challenges for broadcasters operating under impartiality obligations.
Major stations explored alternatives to the stopwatch balance, and complemented their coverage with broader context, explanation, and fact-checking. NHK uploaded unedited footage and full transcripts of party leaders’ campaign-launch speeches and analysed frequently used words. Some commercial broadcasters used ‘pre-bunking’ to limit the spread of misinformation.
Yasuomi Sawa
Journalist and Professor of Journalism, Waseda University
Reiko Saisho
NHK, Broadcasting Culture Research Institute
In the online survey we ask respondents which news brands they have used to access news in the last week. These figures are based on respondents’ recall of the news sources they have used, and should be understood as survey-based measures of weekly brand reach for news.
They are not the same as web analytics, audience ratings, or other audience measurement systems (such as BARB for television in the UK). Those approaches use different methods and may measure different things. Our figures are based on what respondents tell us in an online survey about which brands they have used for news in the past week.
It is also important to note that we ask specifically about use for news. Some multi-genre broadcasters, newspapers, or other providers offer content beyond news, and the figures in our report should not be interpreted as measuring the overall audience reach for these media organisations. They refer only to respondents who say they used that brand to access news.
How do you ask about offline and online news reach?
We ask about offline and online reach separately. First, we ask respondents which brands they have used to access news offline in the last week, via TV, radio, print, and other traditional media. Then we ask which brands they have used to access news online in the last week, via websites, apps, social media, and other forms of internet access.
The questions as asked in the survey are:
Which of the following brands have you used to access news offline in the last week (via TV, radio, print, and other traditional media)? Please select all that apply.
Which of the following brands have you used to access news online in the last week (via websites, apps, social media, and other forms of Internet access)? Please select all that apply.
Respondents can select more than one brand in each question. For that reason, the figures do not add up to 100%. They show the proportion of respondents who say they used each brand for news in the last week.
How do you present the weekly news reach data in the report?
On each country or market page, we present the most widely used brands in two charts: one for offline news reach and one for online news reach. The offline chart covers use via TV, radio, print, and other traditional media. The online chart covers use via websites, apps, social media, and other internet-based forms of access.
The figures shown are weekly news reach: the percentage of respondents who say they used that brand for news at least once in the last week.
These figures are useful for comparing the relative news reach of different brands within each market, and for understanding how news use is distributed across offline and online sources. However, they should not be treated as market shares or measures of total time spent with a brand.
How do you choose which brands to ask about?
The brand selection is a strategic sample and not a comprehensive list of all news providers in each market. We consult with country or market experts, review prior years’ Digital News Report data, and draw on other data sources to identify the most widely used brands for news across traditional and online channels.
In some cases, where a news provider operates a number of related news brands, we aggregate these under a single heading. For example, in the UK, reach for the BBC may include use of several BBC news services across different platforms. This is done to give a clearer sense of the overall reach of major news providers, but it means that figures may not always refer to a single programme, website, app, newspaper, or channel.
Because of survey length limitations, we can only ask about a limited number of brands in each market. The charts should not be treated as exhaustive lists of every brand used for news in that market. Due to space limitations, reach charts show up to 16 of the most used brands, though we ask about more in each survey.
How should the offline and online figures be interpreted?
A respondent may use the same brand both offline and online, and may also use several different brands in the same week. For this reason, the offline and online figures should not be added together to calculate a total audience. They are best read as separate indicators of how far particular brands reach people through different forms of access.
As with other survey-based findings, small differences should be interpreted with caution. We are careful not to claim that one brand reaches more people than another, or that a brand’s reach has changed, unless the differences are large enough to be meaningful. Any year on year change of 2 percentage points or lower is not considered statistically significant.
Changing media
Use of traditional and online media has remained largely stable, with the sole exception of AI chatbots, which have increased in adoption not only among younger cohorts but also among individuals aged 35 to 54.
Pay for online news
9%
(-1)
Avoid the news sometimes/often
12%
(+1)
Trust in news overall
41%
(+2)
Global average: 37%
Trust in news, which fell sharply in the previous year in the context of a sex scandal involving a major television network, may be starting to bounce back. NHK remains the most trusted news source of those asked about in the survey.
We ask each respondent to rate a number of popular brands (usually 15 in each country) according to how trustworthy they think each brand’s news output is. We do this on a 0-10 scale, where a score of 0 means that the respondent does not see the brand as trustworthy at all and 10 means that they see the brand as completely trustworthy – with 5 meaning ‘neither trustworthy or untrustworthy’. There is an option for those who have not heard of any particular brand to ensure that scores are based only on responses from people who are familiar with each brand.
When we come to report these scores, we add up the proportion of respondents that give a score between 6-10 and mark this as ‘trust’. We also add up the proportion that give a brand a score between 0-4 and mark this as ‘don’t trust’.
The question as asked in the survey is…
How trustworthy would you say news from the following brands is? Please use the scale below, where 0 is ‘not at all trustworthy’ and 10 is ‘completely trustworthy’.
As we make explicit throughout the report, including next to tables presenting brand-level trust findings, whether respondents consider a brand trustworthy or not is their subjective judgement. The percentage figures shown are aggregates of people’s personal opinions, they are not an objective assessment of underlying trustworthiness. We leave it to each respondent to form an opinion on whether they trust someone or something, and we field the question because we consider the resulting data to be important.
How do you present the trust data in the report?
We present the data in an alphabetised table. In the past, we presented this data as a stacked bar chart, but this led some to treat the chart as a list of the most and least trusted news brands in a given market, despite our explicit explanation this was not what the tables showed.
We present the data in a way that avoids giving small differences the appearance of great importance. In cases where there is around two percentage points difference or less between the brands, we cannot say for sure that one brand is more trusted than another. We are careful not to try to claim that one brand is more trusted than another or that trust scores have changed unless those changes are statistically significant.
Due to survey length limitations, it is important to note that we only ask about 15 of the most widely used brands. It is very likely that there are brands with lower (and higher) trust scores that we do not ask about. For that reason, we cannot say that any brand is the least (or most) trusted overall. Next to each chart we are careful to say: “Only the below brands were included in the survey. It should not be treated as a list of the most or least trusted brands as it is not exhaustive.”
How do you choose which brands to ask about?
We consult with country or market experts, draw on prior years’ Digital News Report data and other data sources to determine the most widely used brands (across traditional and online channels) when it comes to news. We also try to include ‘local newspapers’ or ‘local television’ as catch-all titles as we recognise their impact is considerable in most markets.
How representative is this 48-market survey?
The Digital News Report survey is based on an online poll but the methodology selects participants to be as representative of national populations as possible. Samples are assembled using representative quotas for age, gender, and region in every market and data is weighted to targets based on census/industry accepted data. The full methodology can be found here.
How do you try to contextualise the findings to ensure that trust scores are not taken out of context or misinterpreted?
Trust is one of a number of measures we track, including consumption of different sources, device usage, social media use, and much more. We aim to maintain consistency in our measurements year-on-year so that ratings of trust, levels of news consumption, and more, can be contextualised.
Country data is accompanied by an 800-word commentary from a media expert that aims to set the data in a wider context. We write a short commentary on the trust scores where appropriate, noting any statistically relevant changes.
RSF World Press Freedom Index
62/180
Score 62.9
Measure of press freedom from NGO Reporters Without Borders based on expert assessment. More at rsf.org
