Surveillance capabilities
AI models and systems may grant governments or corporations increased monitoring over individuals.
ENTITY
2 - AI
INTENT
1 - Intentional
TIMING
2 - Post-deployment
Risk ID
mit1086
Domain lineage
4. Malicious Actors & Misuse
4.1 > Disinformation, surveillance, and influence at scale
Mitigation strategy
1. Establish and enforce stringent legal and regulatory frameworks that explicitly prohibit indiscriminate AI-driven mass surveillance and require mandatory, public transparency reports detailing the purpose, data sources, and algorithmic decision-making parameters of any system employed for individual monitoring. 2. Mandate the integration of Privacy-Preserving Technologies (PPTs), such as differential privacy and secure federated learning, during the design and development phases to ensure data minimization and anonymization, thereby reducing the intrinsic capacity of AI models to create detailed and revealing behavioral profiles of individuals. 3. Implement a continuous, independent auditing and governance structure that requires regular fairness and ethical impact assessments, alongside the maintenance of immutable audit trails, to verify compliance with non-discrimination laws and provide post-deployment accountability for surveillance-related outcomes.