Guidance on how to select, test and embed AI effectively within policing
This guidance is aimed at project leads who are overseeing projects or programmes that use tools or systems enabled by artificial intelligence (AI) to improve police performance and productivity. It provides practical guidance on how to select, test and embed AI effectively within policing. It does not introduce new obligations but should be read in conjunction with the Data-driven technologies authorised professional practice (APP).
For the purposes of this guidance, ‘AI’ refers to technologies with the ability to perform intellectual processes that would otherwise require human intelligence, such as visual perception, speech recognition and language translation.
AI has the potential to transform policing, improving the speed and accuracy of policing activities and enabling officers to focus more on the front line. This guidance aims to support police forces in navigating the challenges and risks that can arise with this rapidly evolving technology so that they can harness its benefits.
The guidance should also be read in conjunction with:
- the Data ethics APP
- the Responsible AI checklist
- the National Police Chiefs’ Council (NPCC) AI Playbook, AI Covenant and AI strategy
Concerns about the potential harms that arise from misuse of AI or poorly procured AI are addressed throughout. The guidance itself is covered by the comprehensive equality impact assessment that was carried out for the over-arching Data-driven technologies APP. The College has produced e-learning on AI fundamentals (you will need to log in to College Learn), which covers general principles, the use of AI to improve productivity and prevent crime and AI as a criminal threat.