Nine out of ten Swedes believe AI-generated fake images could influence the outcome of the country’s upcoming general election in September 2026, according to a new report from the Internet Foundation. The findings highlight growing public concern about artificial intelligence’s role in shaping political discourse and voter perception ahead of what analysts are calling Sweden’s first “AI election.”
The report, titled Swedes and the Internet: Election Special 2026, reveals that 88 percent of respondents are worried that AI will be used to create false images designed to sway election results. Four in ten Swedes admitted they struggle to determine whether a political message online is genuinely human-made or generated by artificial intelligence. These figures underscore a widespread erosion of trust in digital content during a politically sensitive period.
Concerns extend beyond image manipulation. One in five respondents said they have refrained from sharing political views online due to fear of harassment or threats—a statistic that has remained largely unchanged since 2022. Researchers warn that the proliferation of AI-generated content risks fragmenting society’s shared understanding of reality, a cornerstone of democratic decision-making.
“For democracy to function, we must share basic facts and exist in a similar reality,” said Simon Lindgren, Professor of Sociology at Umeå University, during a January 2026 seminar in the Swedish Parliament. “AI-generated content can fragment our shared reality and craft informed decisions harder.” The seminar, organized by Rifo and WASP HS, brought together academics and lawmakers to examine how AI is altering the conditions for democratic discourse in Sweden.
Malin Rönnblom, Professor of Political Science at Karlstad University, emphasized that democracy relies on a collective societal project. “Democracy is living together—a joint project that never ends,” she stated. However, she cautioned that AI’s increasing presence in information flows threatens to undermine this shared experience by amplifying disinformation and eroding consensus on fundamental facts.
The Internet Foundation’s report also notes that while voter guides and party leader debates remain the most influential factors in Swedes’ voting decisions, the rise of AI-driven disinformation poses a novel challenge to electoral integrity. With the general election scheduled for September 14, 2026, authorities and civil society groups are under pressure to strengthen media literacy initiatives and regulatory oversight of AI use in political communication.
As Sweden prepares for what may be the first national election significantly shaped by artificial intelligence, experts stress the importance of transparency, public education, and institutional safeguards to protect the integrity of the democratic process. The coming months will be critical in determining whether Sweden can navigate these technological challenges while preserving public trust in its electoral system.