This essay examines how artificial intelligence has reshaped traditional surveillance into predictive and biometric systems of social control. Drawing on examples from authoritarian and democratic contexts, it considers the ethical challenges posed by algorithmic bias and autonomous weapons before assessing regulatory responses. The discussion highlights the tension between technological efficiency and fundamental rights within civic education frameworks.
A. Le modèle autoritaire : l’exemple de la Chine
China illustrates systematic state surveillance through widespread facial recognition networks often described as ‘Skynet’. These systems integrate with the social credit framework, where algorithms assess citizen behaviour and impose restrictions on individuals deemed non-compliant. In Xinjiang, targeted algorithms profile Uyghur populations, enabling monitoring that raises serious human rights concerns. Such applications demonstrate how AI can consolidate state power with limited accountability.
B. Le modèle démocratique : la police aux États-Unis
In democratic settings, predictive policing tools such as PredPol and platforms developed by Palantir analyse data to forecast crime hotspots. While intended to improve resource allocation, these methods frequently intensify scrutiny of disadvantaged neighbourhoods. Facial recognition databases, including Clearview AI, are used by law enforcement to identify suspects; however, the absence of comprehensive legal safeguards creates risks of misuse. This suggests that even in open societies, AI surveillance can erode privacy without adequate oversight.
C. Biais algorithmiques et discrimination raciale
Algorithmic bias arises when systems are trained on historically skewed datasets, resulting in disproportionate targeting of Black and Hispanic communities. Facial recognition exhibits notably higher error rates for individuals with darker skin tones, contributing to wrongful arrests documented in several US cases. These patterns indicate that without deliberate correction, AI tools may reproduce existing social inequalities rather than neutralise them.
II. L’IA sur le champ de bataille : vers une guerre automatisée et déshumanisée
The migration of civilian AI technologies into military applications raises profound moral questions. Recent adjustments by companies including OpenAI and Anthropic have permitted defence collaborations, while Palantir systems have supported data-driven targeting in Ukraine and the Middle East. Proponents argue that AI accelerates the OODA decision cycle and enables precise strikes that minimise collateral damage. Yet the ‘black box’ nature of many algorithms complicates responsibility for erroneous targeting, and autonomous lethal systems struggle to replicate the contextual judgement required by international humanitarian law.
III. Réglementer l’inacceptable : les défis de l’encadrement juridique
The EU AI Act represents a pioneering attempt to restrict real-time biometric surveillance in public spaces, classifying law-enforcement applications as high-risk and subjecting them to mandatory audits. Internationally, talks at the United Nations on banning lethal autonomous weapons have stalled, largely because major powers cite national security needs. Civil society has responded through employee protests, such as those surrounding Google’s Project Maven, and campaigns by organisations advocating for clear prohibitions. These efforts underscore the difficulty of achieving binding global standards amid competitive pressures.
Conclusion
AI-driven surveillance and warfare present complex challenges that demand both technical safeguards and robust legal frameworks. While regulatory initiatives offer partial remedies, persistent gaps in accountability and international consensus suggest that civic vigilance remains essential to prevent dehumanising applications of these technologies.
References
- European Commission. (2024) The EU Artificial Intelligence Act. European Union.
- Human Rights Watch. (2023) World Report 2023: China. Human Rights Watch.
- Stop Killer Robots. (2023) Campaign to Stop Killer Robots. Stop Killer Robots Coalition.
