The NPCC’s AI Covenant establishes a formal framework to ensure that the rapid adoption of machine learning and advanced analytics across UK police forces remains ethical and legally compliant. It defines AI as technology that can learn, generalize, or infer meaning to match or exceed human performance, currently utilized in areas like 999 call triage, facial recognition, and reoffending risk assessments. By endorsing these principles, police leaders commit to balancing the operational “step change” in efficiency provided by algorithms against the need to mitigate unintended biases and protect civil liberties.+4

To maintain public trust, the framework dictates that all AI use must be lawful, explainable, and subject to “Maximum Transparency by Default”. A critical safeguard within the document is the requirement for a human to remain the ultimate decision-maker for any AI output affecting the public, alongside the necessity for “kill switches” to stop systems that display undesired behavior. Under this governance, Chief Constables are held accountable for ensuring that AI tools are robust, reliable, and used only for their authorized purposes.

Read the article >