By Michael Maher (Researcher, Institute for Social Innovation and Impact, Northampton University and PROBabLE Futures senior researcher) and Marion Oswald (Professor of Law, Northumbria University, PROBabLE Futures PI)
What happens when AI removes the everyday routines that created modern police culture in the first place?
Many of us will remember the iconic 2006 TV series ‘Life on Mars’ in which Manchester police detective Sam Tyler wakes up after a car accident to find he has time-travelled back to the year 1973. There he is surrounded by a police culture of sexism, racism, homophobia and corruption, with little attention paid to suspect’s rights. The police team is led by anti-hero DCI Gene Hunt, a man who although committed to seeing justice done, would rather follow his gut or a hunch than bother with the evidence.
Sam was a fish-out-of-water because he came from the post-PACE generation of police officers. The Police and Criminal Evidence Act 1984, and the later Criminal Procedure and Investigations Act 1996, were introduced in response to the growing concern around police culture and practices in the 1970s and early 80s, resulting in miscarriages of justice and failed investigations. PACE introduced safeguards designed to restore public confidence: new rules on evidence handling, how interviews were conducted, and the practice of stop-and-search. Importantly, it codified the requirement of the ‘reasonable grounds’ officers needed to believe before a more coercive step such as stop-and-search or arrest was taken. The CPIA introduced the disclosure regime by which the prosecution provides the defence with material from the investigation that could undermine the prosecution case or assist the defence, even where that material is not used as part of the prosecution evidence at trial, embedding a ‘thinking’ approach where the prosecution had to highlight potentially useful information, rather than a purely administrative one with generic descriptions (UK Government, 2026).
For decades, PACE and CPIA provided the bedrock of policing practice, not necessarily guaranteeing legal and ethical behaviour, but at least setting the standard for procedural fairness, accountability and the requirement for actions to be explainable, therefore, underpinning the policing-by-consent model. The influence of the two Acts, however, is not simply due to their ascension into law, but their impact on the organisational culture of the police service as it sought to evolve itself into its modern guise. At first glance, AI appears to support the values introduced by PACE: fast investigations, better evidence handling, and less paperwork. However, it may also be quietly changing the culture that PACE helped create.
Organisational theorist Edgar Shein argued that culture isn’t just about values; it’s reinforced every day by routines, paperwork, languages and stories. Over time these visible practices shape what people come to see as normal. In Life on Mars, these are most clearly reflected in the disregard for protocol, Gene’s interactions with suspects, and the physical environment of the police station. Although Sam (and the audience) enter as outsiders, there is an immediate recognition of the police force’s distinctive culture. These lead to espoused beliefs and values that serve a normative moral function by other officers on how to respond to situations, encapsulated by Gene’s assertion that, “The public don’t give a damn what we do, as long as we get results”. Finally, these evolve into basic assumptions that have been socially validated through shared experience and continually reinforced by leaders until they become taken-for-granted, non-discussable assumptions.
Organisational culture, however, is not dependent on bombastic personalities and gung-ho interpretations of policy. The implementation of PACE and the CPIA fundamentally reshaped police culture by introducing greater bureaucracy and administration, enacting formal codes of practice, codifying suspect rights, and reinforcing the legal boundaries of the police. These reforms created new artifacts that an outsider might observe as ‘red tape’, but which serve as a constant small reminder to officers as their role as professional servants of the Crown. In turn, these artifacts helped foster new beliefs and values promoted by leaders, gradually changing the basic assumptions about what it meant to be a police officer.
The success of PACE and the CPIA and the transformation of the Police Force into a Police Service, was scaffolded by an evolution in organisational culture that allowed these changes in perspective to take root. This scaffolding, however, is precisely where the adoption of AI is beginning to take hold and potentially change or even undermine. Police leaders and other stakeholders have promoted AI as a means of relieving officers of the administrative burden that keeps them away from the street and away from what is often framed as ‘real police work’. Yet little attention has been paid to what happens to organisational culture when this scaffolding is gradually eroded, or whether policing is entering a post-PACE era in which a new organisational culture is beginning to evolve.
The rapid expansion of probabilistic AI, including predictive tools, live facial recognition, and large language models, could be contributing to the emergence of institutional voids as incumbent leaders struggle, or are unwilling, to implement policies that mitigate harmful behaviours or uphold the importance of the existing organisational culture and legal framework. This reflects the legal-institutional context in which increasing the speed of crime resolution and improving operational efficiency are prioritised over governance processes. As a result, organisational culture may increasingly be shaped by a new generation of police officers, many of whom, as noted at the recent PoliceAI launch, expressed a desire for a faster, more modern systems. If these technologies deliver a police service that is both quicker and more operationally efficient, there is unlikely to be significant pressure for officers to slow down or question the cultural assumptions that emerge alongside with them. As administrative and bureaucratic processes are moved into the sphere of AI, there is a risk that the professionalisation of the police service circles back around to the Life on Mars ‘streetcraft’ and ‘hunch’ based decisions, this time mediated by technology.
The artifacts of change are already being noticed by outsiders, with the public now aware of technology such as Live Facial Recognition leading to wrongful arrests and the hallucinations caused by generative AI leading to football fans being halted from entering the country. Currently, we are terming the potentially emerging culture as ‘vibe policing’. As vibe coding refers to building software by describing what is required in natural language, whilst AI generates the code, we posture that ‘vibe policing’, as an organisational culture, may lead to the return of hunches, with them being legitimised by the savvy use of AI technology. The continuing use of technology will enviably lead to a new organisational culture, but the mediation of the risks can still take place. The point is not that AI will turn each police officer into Gene Hunt. It is that organisational cultures evolve through everyday practices. If AI removes the routines that reinforced procedural fairness for forty years, it will also reshape the assumptions that underpin modern policing. We would suggest that it is not about whether policing will develop a new culture, but rather if we notice before it begins to feel normal.
For more information on RAi UK’s Keystone project PROBabLE Futures:
PROBabLE Futures: Probabilistic AI Systems in Law Enforcement Futures