Scoping review of frameworks for effective participatory oversight of AI in Law Enforcement

Rapid advancements in Artificial Intelligence (AI) have led to a widespread adoption of AI technologies across critical sectors like healthcare, criminal justice and travel which has revolutionised how these fields meet complex demands and serve diverse needs (Pawar and Khose, 2024; Akter, 2024). AI deployment in law enforcement and criminal justice holds the potential to enhance the use of data for crime prevention and detection whilst improving the way law enforcement and criminal justice work (Oswald et al., 2024). Despite the benefits, the adoption and deployment of AI poses significant risks of interference with fundamental rights and disproportionality (Oswald et al., 2024). Algorithms commonly used in practice can have errors and biases, resulting in decisions based on flawed information and consequently, deepening existing inequalities, creating poor practices, or embedding ineffective strategies (Aquino, 2023). Thus, relying solely on the rigid, rule-bound logic of algorithms for decision-making conflicts with the principle that actions should be responsive to individuals' unique circumstances (Green, 2022).