In the contemporary debate on artificial intelligence, there exists a fault line that few commentators openly acknowledge, yet it is shaping the digital geopolitics with the same intensity as a global financial crisis reshapes markets: AI is not merely a tool for economic and social progress, but increasingly a systematic instrument of repression wielded by authoritarian states, fragile regimes, and even democracies under stress. Anyone who has spent significant time in the global tech ecosystem knows this is not apocalyptic speculation but a reality already underway, with documented cases of algorithms designed, purchased, or adapted to consolidate political power, suppress opposition, and surveil entire populations with a granularity that would have unsettled even Shakespeare. Understanding this phenomenon requires looking beyond the slogans of “inclusive progress” and engaging with the data, governance mechanisms, and legal implications that emerge when artificial intelligence becomes an arm of social control rather than a tool for human empowerment.
When discussing AI as a tool of repression, we are not talking about some distant future, but active practices in countries such as China, where facial recognition systems and predictive behavior models are integrated into surveillance networks monitoring ethnic minorities and political dissidents with precision previously unimaginable. This form of “algorithmic authoritarianism” is not theory; social credit scoring, intelligent sensor networks, and automated early-warning systems for “antisocial behavior” are deployed alongside legislative frameworks legitimizing the use of personal data for public order. From the perspective of a technologist accustomed to evaluating complex platforms, these systems are not mere efficiency tools, but architectures of power, designed to transform citizens into objects of analysis, prediction, and ultimately, punishment. The implications for human rights are profound: privacy becomes an illusion, freedom of expression a luxury, and the very possibility of political dissent is technically eroded by infrastructures that leave no margin for human error.
In this context, the European Union has sought to position itself as a guardian of an alternative vision, embodied in initiatives such as the AI Act, which seeks to regulate AI use by defining risk categories and imposing requirements for transparency and human oversight. However, any seasoned policymaker or business leader knows that a law is only as effective as the will to enforce it; without robust enforcement mechanisms and international cooperation, even the most sophisticated rules risk remaining mere paper. The challenge for the European Parliament is not purely technical or regulatory, but deeply political: how to balance the adoption of AI technologies that drive industrial innovation with the need to protect fundamental human rights in a global landscape where authoritarian actors ignore both ethical standards and international legal obligations.
Examining global cases of abuse requires widening the lens beyond China. In many post-Soviet republics and certain Middle Eastern nations, AI surveillance technologies have been imported alongside commercial contracts and technical assistance agreements, often without public debate or human rights impact assessments. In Egypt, for instance, facial recognition systems have been used to identify and arrest activists and journalists, relying on datasets collected from public cameras and social media platforms. Similarly, in Russia, predictive software has been integrated into protest management to forecast demonstrator locations and preempt mass gatherings. Documented by human rights organizations and international think tanks, these examples show a troubling trend: the adoption of advanced AI tools is not neutral but reflects the political intentions of those in power, legitimizing mass surveillance practices that would have been unthinkable without extensive and intrusive policing infrastructure.
Beyond surveillance, another front of AI abuse concerns automated decision-making systems used in judicial, immigration, and social service contexts. Here, the danger is not merely that the algorithm errs, but that biases embedded in the models perpetuate discriminatory practices. These biases are not accidental but reflect historical datasets and human prejudices, amplified by inadequately supervised machine learning models. In the United States and United Kingdom, risk-assessment systems have faced legal challenges for producing unfavorable outcomes for already marginalized groups, relying on historical correlations that fail to account for structural injustice. These examples underscore a uncomfortable truth: technology does not exist in a moral vacuum; when integrated into public decision-making processes without safeguards, it can reproduce and amplify existing inequalities, transforming the promise of algorithmic fairness into a dystopian parody.
In an era where even established democracies show signs of weakening under pressures of national security and public order, the temptation to deploy AI surveillance solutions cannot be underestimated. Europe itself is not immune: pilot projects for video analysis in train stations, facial recognition to identify “at-risk” individuals, and automated border control systems raise profound questions about the boundaries between security and liberty. Even when justified as protecting citizens, the lack of transparency and public participation in implementation erodes institutional trust and sets precedents for less benign future uses. A democracy that sacrifices fundamental rights on the altar of technological efficiency risks transforming into something irreversibly different from what constitutions and international treaties envisioned.
The international regulatory landscape is equally fragmented. While instruments such as the European Convention on Human Rights or the UN Guiding Principles on Business and Human Rights exist, there is no binding global framework specifically addressing AI use in repressive contexts. This regulatory gap allows states with weak safeguards to adopt invasive technologies without real consequences, while maintaining a veneer of legitimacy on the international stage. This vacuum presents an opportunity for the EU to exercise leadership, promoting rights-based standards and incentivizing regulatory convergence, but doing so requires coherence, resources, and a willingness to engage with global partners often hostile to restrictive criteria.
The diplomatic dimension cannot be overlooked: relations between geopolitical blocs increasingly involve technology agreements, regulatory exchanges, and sanctions linked to AI misuse. The United States, despite internal contradictions, has imposed sanctions on entities supplying surveillance technologies to repressive regimes, signaling that technology is now a terrain of geopolitical contestation. At the same time, G20 nations are discussing guidelines for ethical AI use, though these discussions often remain at the principle level without translating into binding obligations. Without robust technological diplomacy, the risk is the emergence of two divergent orders: one based on rights, transparency, and human oversight; the other on algorithmic surveillance, opacity, and concentration of political power.
Addressing these issues requires multifaceted action: investing in technical capabilities to analyze and counter authoritarian control systems, industrial cooperation to develop human-rights-compliant AI technologies, and strategic use of tools such as sanctions and export controls to limit the proliferation of abusive technologies. It is crucial for those designing, distributing, and using AI systems to understand that technological neutrality is a myth: any algorithm deployed in a power context has material effects on people’s lives. Ignoring this reality is a luxury democracies can no longer afford, especially when AI is at the heart of critical infrastructure and decision-making processes affecting millions.
Ultimately, the weaponization of AI is not merely a technical or legal problem but a symptom of deeper tensions between digital modernization and human rights. The international community is called not only to react but to define an agenda that balances technological ambition with the preservation of fundamental values essential to any free society. For those of us who have spent decades in technology, we know that innovation is a double-edged sword: it can create extraordinary opportunities or unprecedented instruments of control. Institutions, businesses, and civil society must decide on which side of the blade they stand before technology decides for them.
A.D.
