AI in Policing: Human Bias, Moral AI, and the Future of Criminal Investigation by Dr Ivar Fahsing

Fill out the form to get access to the whitepaper.

This whitepaper examines one of the most consequential questions facing criminal justice today: AI in Policing and how AI can support investigative work without amplifying the very biases that lead to wrongful convictions.

Written by Dr Ivar Fahsing, Chief Scientific Advisor at Davidhorn, it draws on more than fifty years of cognitive psychology — from Tversky and Kahneman to the latest research on AI and human decision-making — to explain why the human mind struggles under the conditions that define serious investigations, and what properly governed AI must do to help.

The information explosion has not broadened investigative minds; it has narrowed them. Faced with millions of digital artefacts, rising caseloads, and relentless time pressure, investigators fall back on mental shortcuts precisely when clear thinking matters most. AI could make this worse — or it could become the cognitive infrastructure that keeps alternatives in play, surfaces meaning, and protects the presumption of innocence.

This whitepaper is grounded in decades of peer-reviewed research on investigative decision-making, alongside recent studies from Boston Consulting Group on AI-related cognitive fatigue and from Harvard Kennedy School’s Allen Lab for Democracy Renovation on the moral reasoning of leading AI models.

From confirmation bias and “AI brain fry” to the difference between AI as a bureaucratic amplifier and AI as genuine cognitive offload, this whitepaper explores what it takes to design and deploy AI in Policing that serves the rule of law rather than eroding it.

In this whitepaper, you can learn:

• Why human cognitive bias is the dominant failure mode in investigative work — and why intelligence and experience offer no protection against it
• How the information explosion turned a slow-burning problem into a full-blown crisis
• The difference between AI that amplifies bias and AI that supports sound human judgment
• Which human-in-command safeguards — transparency, evidence traceability, decision logs — must be in place before AI enters high-stakes investigations

Used well, AI finally offers tools that match the scale of the challenge. Used carelessly, it risks becoming the most persuasive confirmation-bias machine ever built.

Understanding how technology can strengthen — rather than quietly erode — investigative reasoning is essential for investigators, police leaders, and anyone shaping the future of criminal justice.