information is accessed and interpreted. The sharpest rises in the use of social media for news consumption have been in the United States, Latin America, Africa and some South-Eastern Asian countries.25 In the United States, the share of people who cite social media as their primary source of news has grown sharply, from 4% in 2015 to 34% in 2025. For the first time, more people in the United States now access news through social media and video platforms than through television or traditional news websites.26 In addition, the use of AI tools for finding information is also rising, from 11% in 2024 to 24% today.27 The Reuters Institute survey also reveals concerns among the general public that AI will make the news less transparent, less accurate and significantly less trustworthy.28 A particular problem area is the proliferation of deepfakes (digitally altered videos, images, and audio recordings). Over the past five years, deepfake creation has become easier, cheaper, and more convincing.29 While the use of deepfakes during the 2024 “super election year”30 was still a relatively new phenomenon, they have started to proliferate and have a greater influence on politics and electoral processes. The weaponization of deepfakes can undermine trust in democratic institutions, contributing to more political polarization, and can lead to the incitement of political violence or social upheaval. Recent elections in the United States, Ireland, the Netherlands, Pakistan, Japan, India and Argentina have all had to contend with such fabricated content on social media, depicting fictional events or discrediting political candidates, blurring the line between fact and fiction.31 As AI is used to make such content more personalized and persuasive, there is a risk of greater impact on elections.32 For example, research has found that 87% of people in the United Kingdom are concerned about deepfakes affecting election results. But while awareness is high, many lack confidence in their abilities to identify when content is manipulated.33 Increasing reliance on both social media and AI tools enhances the impact of algorithmic bias, which shapes what information users see online and reinforces exposure of individuals to information aligned with their views. This can create widely divergent perspectives on real-world events and developments. The impacts are starting to run even deeper. How real-world events are interpreted online combined with the growing circulation of violent content on social media may be leading citizens to become more emotionally and cognitively detached and numbed to human tragedies. There were 61 conflicts across 36 countries in 2024, making it the fourth-most deadly year since the Cold War ended in 1989.34 With content about these conflicts increasingly distributed through algorithms, different perspectives are shared with selected audiences, contributing to a hardening of views. Additionally, repetition of violent content being shared can over time lead to viewers perceiving it as “normal”, generating apathy and disinterest. Studies have shown that exposure to high levels of violent content is linked to emotional desensitization.35 In other words, the way people increasingly consume news and analysis, coupled with the nature of that content, is leading to a disconnect from empathy for other human lives. Jason Leung, Unsplash Global Risks Report 2026 36