When our previous election special was released in 2022, Sweden was at the end of a turbulent term marked by the global COVID-19 pandemic. Shortly after our data collection began, Russia launched its full-scale invasion of Ukraine, and the prevailing security situation dominated the public debate leading up to the election.

Now we are approaching the 2026 election, and the world has changed once again. The recent years have been marked by geopolitical unrest, crises, economic fluctuations, and a rapidly changing information landscape. The conflicts in Ukraine and Gaza and a new political landscape in the U.S., have left a clear imprint on public discourse.

At the same time, generative AI has become a widely available tool. New EU legislation has reshaped the playing field for digital platforms, and issues surrounding disinformation and political influence have gained greater prominence. In this complex world, the voters of today navigate, access information, and form their opinions.

Insights for the 2026 Election Year

In this report, we focus specifically on the digital flow of information during an election year. We examine how political messages are spread online, the nature of the political discourse, and how different groups of voters form their perceptions.

We also examine the role AI plays — and may come to play — in the election. Many Swedes express concern about AI-generated disinformation and manipulation. At the same time, the technology is also used as a tool for understanding politics.

Traditional media continues to have significant influence, but social media platforms like Tiktok play a major and increasingly important role, particularly among first-time voters. At the same time, Facebook is rapidly losing ground as a political arena. New formats are emerging, where both short videos and longer in-depth content, such as podcasts, contribute to how politics is consumed and understood.

This report provides a view of the information landscape in a rapidly changing world. To understand the democratic processes of today and tomorrow, we need to understand how digitalization affects the information we encounter, and the conversations we participate in. This report is a contribution to that understanding. We offer increased insight into Swedes behaviors, attitudes, and actions in the political context online. We want to highlight the internet’s important role in a democratic society.

Below is a short summary of each chapter.

Chapter 1: Political Information and Messages Online

In this chapter, we examine which channels Swedes consider important for accessing political content, and where they typically do so.

Television, radio, and daily newspapers continue to be the channels that Swedes themselves consider most important for political information. This is also where most Swedes access political content.

At the same time, social media are very important platforms for accessing political content, particularly for the younger segment of the population. Social media platforms also enable a different type of content. There, people can follow politicians, influencers, and individual experts own channels, which more than one in four Swedish voters does.

In this year’s survey, we also see that more people are consuming political content via podcasts and on Tiktok. At the same time, fewer people are consuming political information and political messages on Facebook, on TV, and in newspapers.

Traditional media continues to be the most important source of political information among Swedish voters

Television, followed by radio and newspapers, are the channels that most Swedes consider important sources of political information. Traditional media are also the channels that reach the most Swedish voters and reach them most frequently. However, there are significant age differences, with young Swedes obtaining political information to a large extent from other channels, particularly social media.

Half of all Swedes consumes political content on social media

Social media continues to be a major and important platform for political information and political messages. Half of all Swedes engage with political content on social media, and one in four does so daily.

Since October 2025, paid advertising of political content is no longer been permitted on major social media platforms in Europe. This affects both the volume and the type of political content displayed. The proportion of people engaging with politics on social media has decreased slightly compared to 2022. However, the platforms remain major channels for political content, particularly among young people.

Facebook is declining sharply as a platform for political content

Facebook is the platform where the proportion of people engaging with political information and political messages has decreased the most. This is true both in comparison to traditional channels (TV, radio, and newspapers) and other social media platforms. Ahead of the 2022 election, more than 4 in 10 (43%) engaged with political content on Facebook, compared to just under 3 in 10 (29%) who do so in 2026.

There are likely several reasons why Facebook has declined in importance as a channel for political content. The proportion of Swedes who use Facebook frequently (daily or weekly) has decreased slightly since 2022. However, this change is relatively small and cannot on its own explain why fewer people are engaging with political content on Facebook. This suggests that users are also being exposed to less political content on the platform than before. Meta’s new rules prohibiting paid advertising of political content on the platform are likely a key factor in this change. There may also be other explanations, such as changes to algorithms or political actors prioritizing other channels.

More people are consuming political content on podcasts and Tiktok

While Facebook has seen a sharp decline as a platform for political content, podcasts and Tiktok have grown in its place. When looking at the population as a whole, the changes are relatively small. The proportion of people consuming political content via podcasts has increased from 17% to 20%, and on Tiktok, it has risen from 3% to 8% between 2022 and 2026.

Among younger voters, however, the changes are very significant. Among those born in the 2000s, the proportion consuming political content on Tiktok has increased from 22 percent to 38 percent. Podcasts have seen a similar shift among this demographic, rising from 22 percent in 2022 to 30 percent in 2026.

The most significant differences are among this year’s first-time voters. More than twice as many of this year’s first-time voters (48%) consume political content on Tiktok compared to first-time voters in 2022 (23%).

The proportion consuming political content via podcasts has increased from 17% of first-time voters in 2022 to 31% of this year’s first-time voters.

More than one in four Swedes follow politicians, influencers, or experts who discuss politics and social issues

Despite new guidelines on political advertising on major social media platforms, political content remains prevalent for many Swedes. Half of Swedes engage with political content on social media, and more than one in four follow politicians, influencers, or political experts who regularly discuss politics and social issues.

This is particularly common among younger people. For example, nearly 4 in 10 (38%) people born in the 2000s follow individuals who share content about politics and society on social media.

It is also primarily Swedes with a strong interest in politics who follow these accounts, with the proportion of followers being nearly 4 in 10 (37%) compared to only 1 in 10 among people with little interest in politics (9%).

Swedes believe that televised party leader debates and election compasses will be important to them ahead of the 2026 election

About half of Swedes believe that televised party leader debates will be important sources of information when deciding how to vote in the fall 2026 election. Roughly the same number believe that election compasses will be important sources.

Party leader debates on TV are one of the information sources with the potential to reach many voters across generations. In fact, a majority in nearly every generation believes that party leader debates will be important ahead of their vote in the fall of 2026.

Many voters, especially younger people, also believe that election compasses will be important in shaping their decisions. Election compasses stand out because they are expected to be important even for people who are not interested in politics. Many other sources of information tend to reach people with a strong interest in politics to a greater extent, but there are no such differences when it comes specifically to election compasses.

---

Chapter 2: AI in the Political Information Landscape

In this chapter, we show how many Swedes use generative AI services to write about politics and social issues (such as ChatGPT, Gemini, Copilot, and My AI). Among other things, we examine users experiences of writing about politics and social issues using AI services—what topics do they write about? Do they feel the responses are balanced? Do users feel challenged or affirmed in their views on politics and social issues?

In this chapter, we also examine how Swedes feel about different types of regulations for AI services when it comes to politics and social issues, and how many believe that AI will be used to create fake images and posts with the aim of influencing the 2026 election results.

1 in 10 has used AI services to write about politics and social issues

In 2025, about 4 in 10 Swedes had used AI services for some purpose, and this proportion has gradually increased between 2023 and 2025 (see Swedes and the Internet 2025). However, relatively few people have used AI services to discuss politics and social issues during the year.

A larger proportion of men (13%) than women (8%) have used AI services to discuss politics and social issues. It is also significantly more common among people with a strong interest in politics (14%) compared to those with a low interest (6%).

The services are used for everything from simple factual questions to questions requiring more complex answers, such as how one’s own views align with political parties or the pros and cons of various policy proposals. A common use case is asking AI services to describe and summarize how different political parties stand on specific issues. It is also common for young people to use AI services in schoolwork when writing about politics and social issues.

8 out of 10 Swedes believe that AI-generated posts, images, and videos with political content should be clearly labeled

Just over 8 out of 10 Swedes believe that AI-generated posts, images, or videos with political content should be clearly labeled. This is likely closely linked to the fact that many Swedes find it difficult to determine when political content has been created using AI services. More than 4 in 10 find it difficult to determine whether posts or videos with political messages are AI-generated, and only one-sixth (17%) of Swedes find it easy.

Political views and the use of AI services in politics and social issues influence attitudes toward regulation

In the survey, we ask whether Swedes agree that:

• It should be clearly labeled when political and social information (text, images, video) is AI-generated.
• There should be rules governing how political parties and politicians are allowed to use AI.
• There should be rules ensuring that the answers AI provides regarding politics are impartial.

Typically, there are significant age differences in this type of question, but here the differences between younger and older respondents are relatively small. Instead, political views and personal use of AI services in politics and social issues carry greater weight. People who have personally used AI services to write about politics and social issues are generally less supportive of this type of regulation. There are also political differences, with people on the left being significantly more positive to regulation than those on the right.

Few believe that AI will contribute to less polarization in politics and social issues

Expectations that AI services will help increase understanding of political opponents are very low among Swedes.

More than 4 in 10 Swedes believe that AI will reduce people’s understanding of opponents in politics and social issues by providing answers that confirm existing opinions. Only 7 percent believe that AI will increase people’s understanding of opponents by providing answers that open up new ways of thinking.

Even among those who use AI services themselves to write about politics and social issues, confidence is limited. While this group is certainly more positive than the general population, only one in five users believes that AI will increase understanding, and just over 4 in 10 believe that understanding of those with opposing views will decrease because AI services reinforce existing opinions.

Swedes believe AI will be used to create fake images and videos to influence the 2026 election outcome

Nearly all Swedes (88%) expect AI to be used to create fake images or videos to influence the election outcome in the fall of 2026. This proportion is particularly high among men and people with a strong interest in politics. 

---

Chapter 3: Political Discourse Online

This chapter focuses on political discourse online and how the climate of discourse has changed since the 2022 election year. We examine who is participating, what is being discussed, and whether more or fewer people are taking part in the discussions. We also highlight the downside of these conversations and what can hinder open dialogue. How common is it to be subjected to hate and threats when expressing political views online? How many people refrain from participating to avoid being targeted? And does a hostile conversational climate lead to more people refraining from political engagement?

A turbulent world influences the topics discussed online

3 out of 10 Swedes have discussed a political issue in the past year. This is the same proportion as before the 2022 election.

Immigration, migrant integration law, order, and crime prevention remain the political issues most frequently discussed online. At the same time, the turbulent world around us is evident in these conversations. Defense and security policy has become a significantly more common topic of discussion compared to 2022. A larger proportion of people are also discussing foreign policy as well as jobs and employment. It should be noted here that the data collection for "Swedes and the Internet: Election Special 2022" was conducted at the turn of the year 2021–2022, prior to the full-scale invasion of Ukraine.

Who is listening when Swedes discuss these issues?

Although 3 in 10 Swedes have discussed a political issue online in the past year, few—only 5 percent—say they have tried to influence others to change their opinions. This proportion has also decreased slightly compared to 2022, when 8 percent gave the same response. This suggests that the purpose of discussing politics online is rarely to convince others.

Why, then, do Swedes express political views online? One of the most common ways to express political views is by sharing or forwarding posts on social media, something 17 percent have done over the past year. This may indicate that some people see value in displaying or signaling their political stances, even if they do not necessarily expect to influence others opinions.

Just over 1 in 10 people who discuss politics online have been subjected to hate or threats during the past year

Political discussions online are not always constructive. Many Swedes report a hostile online discourse that can both limit and discourage people from participating in political discussions.

Just over one in five (22%) have refrained from expressing their political views online due to the hostile discourse. In total, 8 percent of Swedes report that they have been subjected to harsh criticism, hate, or threats when discussing politics online over the past year. Of these, 4 percent have been subjected to more severe attacks in the form of hate or threats. Among those who have actually participated in political discussions online, the proportion is significantly higher. In this group, a total of 2 in 10 have been subjected to harsh criticism, hate, or threats. Of these, more than 1 in 10 (12%) have been subjected to hate or threats.

A hostile online discourse affects people with views on both the right and the left, but in different ways

People with political views on both the left and the right are affected by a harsh online discourse. The proportion who has been subjected to harsh criticism, hate, or threats is equally high among those who are clearly on the left as among those who are clearly on the right—just over 1 in 10 in the past year.

At the same time, there are other differences between the groups:

• It is primarily people with views clearly on the left (22%) who report posts for harsh language, hate, or threats on social media platforms.
• A larger proportion of people on the far right (29%) compared to those on the far left (21%) have refrained from expressing or discussing political views to avoid harsh criticism, hate, or threats.
• A larger proportion of people with clearly left-wing views (25%), compared to those with clearly right-wing views (14%), have refrained from engaging in politics during the year due to the online discourse climate.

---

Chapter 4: Where Are Opinions Formed? An In-Depth Look at Three Voter Groups

This chapter examines how the political information landscape differs for three distinct voter groups: first-time voters, parents, and retirees. These three voter groups differ significantly from one another in terms of how they access political information, the role digital channels play, and the extent to which they themselves participate in the political conversation online.

The voter groups we are examining consist of the following individuals:

First-time voters use many different channels for political content, but social media is used most frequently

First-time voters differ in many ways from the other two voter groups, parents and retirees. First-time voters access political information and messages through a wide variety of channels. Most access political information in traditional media (TV, radio, or newspapers) from time to time.

Note that "newspapers" here include daily newspapers, evening newspapers, and local newspapers, both online and in print.

Social media, on the other hand, plays a particularly central role, as it is where the majority of first-time voters engage with political content on a daily basis. Instagram, Tiktok, and Youtube are the social media platforms where the most first-time voters engage with political content, with nearly half doing so.

First-time voters are also pioneers when it comes to using AI services to write about politics and social issues, with one in four having done so during the year, compared to only one in ten among the general population.

There are a few additional factors that particularly distinguish how first-time voters engage with and participate in political conversations online:

• First-time voters are largely the only ones engaging with political content on Tiktok. This not only sets them apart from older voters but also specifically distinguishes this year’s new voters. In fact, more than twice as many of the 2026 first-time voters engage with political content on Tiktok (48%) compared to not even one in four (23%) of the first-time voters in 2022. The difference appears to stem partly from the fact that Tiktok is generally used to a greater extent in this group, and partly from the fact that a larger proportion of users actually engage with political content on the platform.
• More than 6 in 10 first-time voters believe that election compasses (65%) and party leader debates on TV (60%) will be important to them ahead of the 2026 election. They share these views with many older voters as well.
• First-time voters opinions appear to be particularly influenced by political discussions online. More than 2 in 10 (22%) have changed their opinion following political discussions online, compared to only 7% of the general population. This proportion is also significantly higher than among first-time voters in 2022, where just under one in seven responded similarly (13%). First-time voters are also more active when it comes to discussing political issues online compared to older voters, with nearly half (47%) having done so during the year.

Parents media habits resemble those of the general population when it comes to engaging with politics

The "parents" voter group is in the prime of life. About 9 out of 10 are between the ages of 26 and 55, but the majority are between 36 and 55. This largely reflects their media habits when it comes to engaging with political information and content. Parents belong to the group that most closely resembles a cross-section of the population. Political interest, however, is generally lower compared to both younger and older voters, which has certain effects on the extent to which they engage with politics online.

Traditional media (TV, radio, or newspapers—both print and online) are used by most parents, and a majority also consume political information daily through these channels. Nearly 6 in 10 also engage with political content on social media, but not nearly as often as first-time voters. A quarter engage with political content daily on social media, primarily on Facebook, followed by Instagram.

Parents are also characterized by:

• Facebook is the largest social media platform for political content. Although Facebook has declined as a channel for political content across all age groups since the previous election year, the platform remains popular among parents (37%) compared to both younger and older voters.
• One in four parents (24%) consumes political content via podcasts. This proportion is not as high as among first-time voters (31%) but higher than the average for the general population (20%) and significantly higher than among retirees (7%).

One way to engage with political content on social media is by following politicians, influencers, or experts who regularly discuss politics and social issues on social media. Nearly one in three (31%) parents does this. This is also one of the few habits where it seems to matter whether one has children living at home. Specifically, a larger proportion of parents with children living at home follow social media accounts with political content (31%) than people in a similar age group without children at home, where one in four responds similarly. Otherwise, the differences are small between parents with children at home and other voters of the same age. This suggests that age, rather than parenthood itself, is the main factor explaining how this voter group engages with politics online.

Retirees frequently engage with political content, primarily from TV, radio, or newspapers

The retiree voter group is a major consumer of politics—nearly all of them engage with political information and content on a daily basis. Unlike first-time voters and parents, their daily political information comes almost exclusively from traditional media such as TV, radio, or newspapers in print or online (91%). Only 13 percent engage with political content on social media daily, and a total of one in four say they do so occasionally (25%). These platforms thus play a significantly smaller role for retirees than for younger voters.

At the same time, a relatively large proportion of retirees (21%) follow politicians, influencers, or experts who discuss politics and social issues on social media.

Retirees discuss politics online to a lesser extent than younger voters. One in four retirees (25%) has discussed a political issue during the year, compared to nearly half of first-time voters (47%). At the same time, retirees constitute a very large voter group. They make up about 29 percent of the voting-age population, while first-time voters make up just under 7 percent. This means that retirees, despite a lower proportion participating in discussions, still account for a relatively large share of the voices in online political conversations.

Overall, this suggests that social media is significant for at least a limited portion of retirees, even though traditional media are clearly the dominant sources.

Retirees also stand out in that:

• More than 6 in 10 retirees believe that party leader debates on TV will be important ahead of their election in the fall of 2026.
• By comparison, few retirees believe that election compasses will be of great importance to them. One in three retirees (33%) believes that election compasses will be important to them, compared to half of parents (56%) and nearly two-thirds of first-time voters (65%).
• Virtually no retirees have started using AI services for politics or social issues; only a few percent have tried them at some point during the year.

Are there any common platforms among the voter groups?

Although the three voter groups (first-time voters, parents, and retirees) differ in many ways when it comes to consuming political content online, there are also some commonalities. Listed below are the channels and sources that reach the majority of voters in all three voter groups:

• Television and newspapers reach a large portion of voters with political information, although first-time voters engage with political content through these channels significantly less often than older voters.
• Radio also reaches many voters, including the youngest. Once again, first-time voters engage with political content on the radio less frequently than older voters, but despite this, half of the new voters engage with political content on the radio to some extent.
• In all three voter groups, a majority believes that party leader debates on TV will be important to them ahead of the 2026 election. Party leader debates thus have the potential to reach both young and old, at least when voters themselves are asked to predict what will be important ahead of their election. 

---

Information about the survey

The Swedish Internet Foundation commissioned the research firm Novus to conduct the data collection, which took place between December 10, 2025, and January 14, 2026. The main study was conducted among the general public aged 18 and older (n=2,417), with a response rate of 57 percent. Data was collected from Novus’s randomly recruited Sweden Panel. The study results are weighted against known population figures (according to Statistics Sweden) with regard to gender, age, education, geography (NUTS2), and political party in the 2022 election.