AI technology has proliferated extensively and offers authoritarian leaders three crucial advantages: scale, speed, and efficiency. As a result, it has undoubtedly become a crucial tool of repression. But AI tools also give individuals and activists the means to organize, resist, and hold their governments accountable. Today, it is far more difficult for autocratic leaders to hide their corrupt practices and govern in darkness.
In 2021 Carnegie launched the Digital Democracy Network and released a first collection of essays that probed the tension between global technological advances and disruptions to democracies.1 Authors discussed the digital consequences of the COVID-19 pandemic, how authoritarians were experimenting with AI systems to fulfill their repressive objectives, and the growing threat from online disinformation. Fast-forward to 2026, and these issues have become sharper and more trenchant. AI and digital technologies have become rooted in political governance strategies. Authoritarian states, such as China, Iran, Russia, and Saudi Arabia, continue to push the boundaries of innovation in their bid to assert state control and limit the ability of citizens to dissent or challenge their authority.
Democracies have not been immune either. As elected leaders have turned against democratic norms and institutions since coming into power, they have embraced tactics that corrode privacy, raise chilling questions about the scope of state surveillance, and leave a shrinking pool of countries with the means to fight back against these trends. The numbers are grim. As the figure below shows, a growing array of countries were implementing surveillance, censorship controls, internet shutdowns, and disinformation strategies in 2025. In 2021, eighty-six countries scored above the mean level of digital repression, with an average score of 0.12. By 2025, that number had risen to ninety-one countries, and the overall mean score had increased to 0.16. Not only have authoritarian states intensified their repression (for example, China, Russia, North Korea, Turkmenistan, and Myanmar), but countries historically on the low end of the repression scale have seen rises as well (for example, the United States, Argentina, and Italy).
What is driving these rises in repression? Undoubtedly, a big factor is the introduction and widespread adoption of AI technologies. Back in 2021, AI was used primarily by a small number of repressive states, with China at the forefront. Today, that is no longer the case; the technology has proliferated extensively. AI offers authoritarian leaders three crucial advantages: scale, speed, and efficiency. It gives security forces the ability to sift through volumes of information at much faster rates than manual methods. Agents can scrutinize personal communications, social media posts, bank records, geolocation data, drone footage, and biometric sensors to derive insights about individuals. As the inference costs associated with advanced AI models plummet, it is increasingly cost-effective for lower-capacity countries to use these instruments. As a result, the world is experiencing the “democratization” of repression, with more and more leaders acquiring and deploying new tools of coercion.
Yet hope remains. These same technologies also give individuals and activists across political systems the means to organize, resist, and hold their governments accountable. Today, it is far more difficult for autocratic leaders to hide their corrupt practices and govern in darkness. In countries like Hungary, which had taken an increasingly autocratic turn, exposing the state’s corrupt machinery and opening new channels of communication beyond state-controlled media contributed to shifting the electoral tide and unseating a powerful incumbent.2 Similarly, in Nepal and Bangladesh, Gen Z protesters armed with smartphones and social media accounts mobilized popular outrage against nepotism and self-dealing, forcing their respective leaders from office. Even in closed political systems, emerging technologies are helping individuals challenge the power of the state to suppress, deny, and withhold information about how it operates.
To better understand these developments, we have assembled ten essays from members of the Carnegie Endowment’s Digital Democracy Network.
The first tranche of essays explores how digital repression strategies continue to adapt, as well as how individuals are responding. Irene Poetranto’s essay examines how regimes in China, Myanmar, and Russia are exploiting the physical infrastructure of the internet to restrict the flow of information to their citizens. She explains how state authorities have coerced internet service providers and facilitated information controls through the passage of draconian laws and regulations. Mahsa Alimardani analyzes how the Iranian regime is using internet access as a bargaining chip for its political aims. She describes a simple transaction the regime offers users: Amplify state propaganda in exchange for access. This adaptation reflects the regime’s need to address its deepening legitimacy problem.
Janjira Sombatpoonsiri focuses on transnational repression in Thailand. She argues that in regions with neighboring autocracies, transnational repression draws on host states’ domestic infrastructure, which authorities repurpose to target dissidents from nearby countries. She highlights three ways in which Thailand’s domestic infrastructure is being used for transnational repression: tracking the whereabouts of dissidents through CCTV footage, surveilling dissidents’ private communications and locations with interception technologies, and monitoring social media to crack down on dissident gatherings.
Nathan Law probes how Chinese authorities and platforms are increasingly adopting AI-driven censorship tools to suppress free speech. Law finds that while these tools are rapidly proliferating, they come with limits: The more tightly a state represses speech, the less useful data it generates for training algorithms. Frank Cheng-Shan Liu explores how governments are using AI-enabled strategies to shape public opinion and influence collective cognition. He focuses on China’s efforts to impose a simplified “unification versus independence” framework on Taiwan, erasing the complexity of Taiwanese public opinion and strengthening Beijing’s One China narrative abroad. Finally, Usama Khilji explores the implications of Gen Z protests across Asia, focusing on youth-led movements in Nepal and Bangladesh. Khilji finds that demonstrations have three factors in common: frustration with inequality, authoritarian tendencies of elected governments, and the exclusion of young people from policymaking.
A second set of essays examines how AI is affecting the exercise of power, spanning sovereignty debates, China’s normative policy agenda, and the deployment of AI systems in war. Luca Belli’s piece discusses how to operationalize the concept of AI sovereignty in a practical way and proposes that countries adopt an approach that prioritizes agency, interoperability, and openness across the digital stack. Shreya Joshi examines how middle powers are pursuing AI sovereignty, focusing on India and the United Arab Emirates, and argues that “managed interdependence” provides a useful framework for understanding these strategies.
Arindrajit Basu turns his lens to Beijing’s global AI policymaking efforts. He contends that the United States’ retreat from international regulatory spaces under President Donald Trump has created an opening for Beijing to reshape global AI norms and institutions to its advantage.
Finally, Yahli Shereshevsky describes how military AI systems are changing battlefield decisionmaking and expanding the scale of destruction in contemporary war. He focuses on two questions: the extent to which automation bias influences the military’s use of AI targeting systems, and whether military AI systems are eroding the principle of restraint in international humanitarian law by increasing the scale and speed of combat operations.
Together, these essays offer crucial insights into the changing digital and technological landscape and highlight emerging developments that are transforming society and politics.
The Carnegie Endowment for International Peace thanks the Charles Stewart Mott Foundation for its support for the Digital Democracy Network. The authors alone are responsible for the views expressed.
Director and Senior Fellow, Democracy, Conflict, and Governance Program
Steve Feldstein is a senior fellow and director of the Democracy, Conflict, and Governance Program at the Carnegie Endowment for International Peace. His research focuses on technology, national security, the global context for democracy, and U.S. foreign policy.
Shreya Joshi
Former James C. Gaither Junior Fellow, Democracy, Conflict and Governance Program
Shreya Joshi was a James C. Gaither Junior Fellow in the Carnegie Democracy, Conflict and Governance Program.
Carnegie does not take institutional positions on public policy issues; the views represented herein are those of the author(s) and do not necessarily reflect the views of Carnegie, its staff, or its trustees.
The hybrid warfare landscape is evolving rapidly, leaving policymakers without clear strategies. To better inform their work in addressing emerging challenges, governments must dig deeper into the underlying dynamics at play.
Raluca Csernatoni, Alicia Wanless
Autonomous AI agents are increasingly prevalent in cyberspace. The EU needs a real-time monitoring strategy, to invest in AI defenses, and to reduce its strategic dependence on U.S. frontier models.
Raluca Csernatoni, Patryk Pawlak
Defense tech innovations will be at the heart of Europe’s new security strategy. But so far, Brussels has been making moves without a broader plan, undermining readiness and credibility.
Raluca Csernatoni
In Ukraine, Gaza, and Iran, AI warfare has come to dominate, with barely any oversight or accountability. Europe must lead the charge on the responsible use of new military technologies.
Raluca Csernatoni
For years, the United States anchored global cyber diplomacy. As Washington rethinks its leadership role, the launch of the UN’s Cyber Global Mechanism may test how allies adjust their engagement.
Patryk Pawlak, Chris Painter
