Public trust is the bedrock of legitimate, effective law enforcement. Its importance grows when law enforcement agencies adopt AI systems. Public attitudes towards AI in policing remain cautious, and trust in law enforcement agencies can strongly influence whether the public accepts new technologies. Therefore, it is up to law enforcement agencies to act effectively and, above all, fairly when making decisions about whether, when and how to adopt and implement AI systems.

Transparency – through clear, open communication about AI systems and meaningful public engagement – is essential to improving trust in how law enforcement agencies use AI systems, safeguarding human rights, enabling scrutiny and enhancing system quality, as well as encouraging sustainable adoption. Yet transparency is commonly challenged by organizational cultures, operational confidentiality, vendor restrictions and limited resources, among other obstacles.
This report offers comprehensive guidance to help law enforcement decision-makers and related actors build and maintain public confidence, through transparency during the responsible implementation of AI systems.

Key findings and recommendations:
•    Law enforcement agencies should aim to develop and maintain an organization-wide culture of transparency that reflects these values by:
o    Being responsible and trustworthy when introducing AI systems.
o    Investing resources in transparency.
o    Understanding the context and listening to the needs of the public.
o    Being patient and consistent, as trust is not built overnight.
o    Starting to pursue transparency as early as possible and maintaining it throughout the system’s life cycle, rather than reacting only once issues arise.
o    Sharing positive and truthful stories about how AI supports public safety while also being open about challenges and mistakes.
o    Clearly explaining what cannot be shared and why, but also embedding transparency requirements in vendor contracts.
o    Being prepared to respond quickly and thoughtfully to public questions.
o    Continuously improving and adapting.

•    To build and sustain trust, law enforcement agencies should aim to share clear, understandable information about AI initiatives, according to strong, strategic public communication practices, such as:
o    Clearly disclosing the use of AI systems and explaining how they serve the public interest.
o    Tailoring communication to the different types of AI systems, audiences and contexts.
o    Involving multidisciplinary experts in communication efforts.
o    Using clear, accessible language and engaging communication formats across multiple channels to make AI systems understandable and inclusive.
o    Starting communication early, before deployment, and maintaining consistent, proactive outreach throughout the AI system’s life cycle.

•    As trust is a two way process that needs to be informed and earned through meaningful public engagement, grounded in a good-faith commitment to genuine consultations which:
o    Involve the entire law enforcement organization in an effort to be relatable, respectful and accessible to the public.
o    Address a wide range of stakeholders, including technology providers, civil society organizations, academics, intermediaries, mediators and communities directly affected by AI systems. 
o    Pay particular attention to individuals and groups more susceptible to discrimination or being left behind, such as the elderly, children, minorities and people with disabilities. 
o    Use diverse formats and modalities for engagement adapted to each context, such as organized visits to law enforcement agencies, community engagement days and online engagement hubs.

This report is the result of research collaboration with BI Norwegian Business School under the project AI4Citizens: Legal, Ethical, and Societal Considerations of Implementing AI Systems for Anonymized Crowd Monitoring to Improve Public Safety. The work was generously supported by the Norwegian Research Council.