The Politics of AI Surveillance: Who Controls the Digital State?
Since the public launch of large-language models like ChatGPT and OpenAI in 2020, Artificial Intelligence (AI) is gaining ground across a variety of private and public areas, the prospect of not only facilitating mundane tasks but also revolutionising labor markets, research, medicine and militaries.
The gilded age of AI
But as the presence of AI is becoming an increasingly normalized part of everyday life, from summarizing texts, fact-checking a statement or composing an email, it is easy to overlook the more nefarious purposes of surveillance, discrimination and persecution for which AI can be used at the state level. This is an increasingly pertinent issue, with the surge of state-based AI surveillance—such as ’safe cities,’ facial recognition, and smart policing—since 2018, extending to at least 75 of the 175 countries with available data. While this trend is present on all continents, there are regional disparities in application, with AI surveillance present in almost 70% of the surveyed African states, over 50% of South East Asian states, and just under 40% of European countries use AI for surveillance. Thus, AI surveillance is not limited to authoritarian states; according to one report, 51% of liberal democracies use AI for surveillance purposes. How, then, is AI being used for surveillance in China, the Middle East, US, and Europe?
China—a spearhead for surveillance
China dominates the AI surveillance sector, with companies like ZTE and Huawei present in over 63 countries, vastly outnumbering the US. This presence is especially noticeable in Africa and Asia, where the use of Chinese surveillance technology correlates closely with participation in the cross-continental Chinese Belt and Road Initiative. In particular, China has been exporting its ‘safe city’ model, which has already been domestically implemented in cities like Beijing as part of its social credit system, to Saudi Arabia, Uganda, and Thailand as well as European cities like Valenciennes, which in 2017 was gifted safe city technology by Huawei. This model connects an extensive network of facial recognition cameras and police body cameras into intelligent command centers using algorithms to predict crime.
Individual freedom versus national security
While states are justifying these measures by reference to crime reduction and national security, organisations are warning about the implications of AI surveillance for privacy, systemic discrimination civil rights and democratic freedoms as AI allows for cost efficient surveillance at an unprecedented spatial and temporal scale. For example, China has domestically implemented large scale AI surveillance encompassing over 600 million cameras, coupled with large language models for minority languages to sharpen its surveillance of the communication of its Tibetan, Uyghur, Korean, and Mongolian minorities. In the Xinjiang province, the Chinese state has created an Integrated Joint Operations Platform, which employs an extensive network of CCTV cameras, facial recognition devices, and or WiFi surveillance devices to suppress political dissent among the province’s Uyghur minority. Such Chinese technology has reportedly also been exported to Saudi Arabia and Iran for similar purposes of suppressing political dissent, and to enhance the precision of drone air strikes in Ukraine and the Middle East.
AI surveillance beyond autocracies
However, the West is not immune to these developments. The US government recently found itself in a legal dispute with AI company Anthropic after the company refused to allow the government to use its ground breaking AI model Claude for domestic surveillance without built-in restraints. The US government claimed that this jeopardised national security by preventing the state from identifying espionage. In addition, US President Trump has issued various executive orders to increase the adoption of AI by federal agencies over state regulations. Indeed, the US already uses surveillance technology deployed by Israel on the occupied West Bank, to stem migration on the Mexican border. Moreover, the Federal Bureau of Investigation (FBI) admitted in March 2026 that federal agencies are buying personal data from data brokers, including location data collected by private companies, in order to track citizens.
Europe: between security, migration and regulation
Meanwhile, the European Union (EU) is exploring Automated Border Crossing technologies. The intelligent system iBorderCtrl is currently being piloted in Greece, Hungary and Latvia applies AI lie detectors to immigrants, with immigrants found lying being automatically detained for further questioning. This system has been criticised by human rights activists and academics as a scientifically weak and potentially discriminatory practice. Thus, even though AI is more regulated in Europe than elsewhere in the world, with the EU AI Act of 2024 restricting large scale usage from sensitive areas through, the risk of questionable AI use in the name of national security remains salient.
Indeed, several member states are stretching the AI Act’s limitations on large-scale surveillance. For example, Luxembourg has since 2025 pursued plans of expanding its use of Trojan spyware from state security and terrorist threats to encompass a broader range of crimes, such as child exploitation, currency counterfeiting and human trafficking. Similarly, the government of Ireland is seeking to expand the powers of the police and Defense Forces to intercept conversations on encrypted platforms like WhatsApp, and iMessage, and other social media platforms. Meanwhile, the Czech Republic was forced to end its use of facial recognition at Prague Airport after six months as it was found to violate the EU AI Act. Likewise, Hungary authorized the police to use real-time facial recognition to identify participants in LGBTQ+ parades in April last year, in violation of the AI Act.
Digital emancipation or authoritarianism?
Thus, it appears that national and international regulation has been lagging behind the rapid tech innovation of recent years. As with any innovation, AI is a neutral tool—but it can be used in ways good or bad depending on the decisions of power-holders. Thus, the application of AI calls for increased scrutiny, accountability and implementation to safeguard the benefits and prospects of improvement it holds out from being hijacked by nefarious purposes undermining democracy and human rights.