In Europe and around the world, AI systems are developed and deployed for harmful and discriminatory forms of state surveillance. From the use of biometrics for identification, recognition and categorization, to predictive systems in various decision-making and resource allocation capacities, AI in law enforcement disproportionately targets already marginalized communities, undermines legal and procedural rights, and enables mass surveillance.
When AI systems are deployed in contexts of law enforcement, security and migration control (including the policing of social security), the power imbalance between the authorities and the surveilled is even more profound. This means that there is an even greater risk of harm, and violations of fundamental rights and the rule of law.
We point to the specific dangers to freedom of assembly, liberty, the right to asylum, privacy and data protection, the right to social protection, and non-discrimination when such technology is deployed by those authorities.
Civil society organisations are calling for an AI Act that prevents unchecked forms of discriminatory and mass surveillance. In order to uphold human rights and prevent harm from the use of AI in policing, migration control and national security, the EU AI Act must:
- Include legal limits prohibiting AI for uses that pose an unacceptable risk for fundamental rights. This includes a legal prohibition on different forms of biometric surveillance, predictive policing, and harmful uses of AI in the migration context.
- Provide public transparency and oversight when police, migration and national security agencies use ‘high-risk’ AI, by upholding an equal duty of these authorities to register high risk uses in the EU AI database.
- Ensure that the AI Act properly regulates the uses of AI in policing, migration and national security that pose risk to human rights, specifically the full list of AI in migration control, and ensuring that national security is not excluded from scope.