Technology and polarization 1.5 6% 17% 25% 23% 16% 11%Short-term (2 years) risk severity score: Societal polarization FIGURE 1.22 Source World Economic Forum Global RisksPerception Survey 2024-2025. 2-year rank: 4th 2%Present or perceived ideological and cultural divisions within and across communities leading to declining social stability, gridlocks in decision-making, economic disruption and increased political polarization. 2-year average risk severity score: 4.6 Proportion of respondents NoteSeverity was assessed on a 1-7 Likert scale[1 = Low severity, 7 = High severity]. The percentages in the graph may not add up to 100% because values havebeen rounded up/down.2 years 7 High Low 6 5 4 3 2 1Severity – Rising use of digital platforms and a gr owing volume of AI-generated content are making divisive misinformation and disinformation more ubiquitous. – Algorithmic bias could become more common due to political and societal polarization and associated misinformation and disinformation. – Deeper digitalization can make surveillance easier for gover nments, companies and threat actors, and this becomes more of a risk as societies polarize further. An estimated two-thirds of the world’s population – 5.5 billion people46 – is online and over five billion people use social media.47 The increasing ubiquity of sensors, CCTV cameras and biometric scanning,among other tools, is further adding to the digitalfootprint of the average citizen. In parallel, theworld’s computing power is increasing rapidly. 48 This is enabling fast-improving AI and GenAImodels to analyse unstructur ed data more quickly and is reducing the cost to produce content. WithSocietal polarization ranking #4 in the GRPStwo-year ranking, the vulnerabilities associatedwith citizens’ online activities look set to continuedeepening hand in hand with societal and politicaldivisions. Taken as a whole, these developmentsthreaten to fundamentally undermine individuals’trust in information and institutions. Like last year, Misinformation and disinformation tops this year’s GRPS two-year ranking. The amount of false or misleading content to which societies are exposed continues to rise, as does the difficulty that citizens, companies and governments face in distinguishing it from true information. The interplay of rising Misinformation and disinformation with political and Societal polarization creates greater scope for algorithmic bias. If human, institutional and societal biases are not addressed, and/or best practices in modelling are neglected, the conditions will be ripe for algorithmic bias to become more prevalent. Such bias, whether inherent in data, models or their creators, can lead to unjust outcomes.Despite the dangers related to false or misleading content, and the associated risks of algorithmic bias, citizens need to strike a balance between privacy on one hand and increased online personalization and convenience on the other hand. While data governance and regulation vary worldwide, it is becoming easier for citizens to be monitored, enabling governments, technology companies and threat actors to reach deeper into people’s lives. Those with access to rising computing power and the ability to leverage sophisticated AI/GenAI models could, if they choose to, exploit further the vulnerabilities provided by citizens’ online footprints. Rising political and Societal polarization could become more of a driving force for such increased surveillance. Misinformation and disinformation in a polarized world The advent of new technologies and the increase in user-generated content platforms is leading to a corresponding rise in the volume of content online. Flows of Misinformation and disinformation from those creating it are becoming more challenging to detect and remove in an increasingly fragmented media landscape. Differentiating between AI- and human-generated false or misleading content – in the form of video, Global Risks Report 2025 34