With rapid advances in technological development and the availability of artificial intelligence (AI) applications, police forces are increasingly looking to maximise the potential that the new landscape offers.
Successful AI adoption does not begin with the technology, but with problem definition. Police agencies must first ask a fundamental question: What problems can be better addressed by an AI-enabled solution?”
John Kilburn, SAS
However, that landscape brings its own challenges. How can law enforcement agencies fully harness the opportunities available to them – particularly in how they use data – while ensuring that both their communities and their own people have absolute confidence in the technology, the decisions it informs, and the transparency and accountability underpinning its use?
John Kilburn, Law Enforcement and Public Safety Lead with data and AI specialists SAS and a former long-serving officer with Queensland Police Service, is well versed in these challenges, having worked with policing agencies around the world to support their adoption and implementation of AI.
He argues that there are two key elements to ensuring that the use of AI is both effective and ethical; firstly, having a clear understanding of the operational impacts of the technology, and secondly, having robust guardrails and frameworks in place to guarantee that the AI used is explainable, auditable and open to scrutiny.
“Successful AI adoption does not begin with the technology, but with problem definition,” stressed John. “Police agencies must first ask a fundamental question: What problems can be better addressed by an AI-enabled solution?
“It is critical to distinguish whether AI is being used for risk identification, prioritisation, or decision support. Only with this clarity can organisations set realistic expectations and avoid misuse or overreach.”
Equally important, he emphasised, is governance: “Trustworthy AI cannot be retrofitted after deployment. Instead, accountability must be built into the design and maintained throughout the entire data and AI life cycle. Only then can lasting trust be established between law enforcement agencies and the communities they serve.”
Turning data into actionable intelligence
At a practical level, one of AI’s most immediate benefits is helping police forces manage vast and complex volumes of data – not by increasing information, but by making it actionable.
This is using AI to risk score and assist with triaging what’s urgent… It allows agencies to deploy their finite resources appropriately to ensure the greatest level of public safety.”
John Kilburn, SAS
“When we’re looking at benefits, I think one of the main areas is helping police to triage the data they have, utilising AI to identify risk inside that information, and then identifying where police need to prioritize and respond in the most timely and appropriate manner,” said John.
Citing the sentiment analysis of multiple data sources, he explained: “For example, from a child protection perspective, you may have one conversation that says ‘I’m looking forward to the weekend – I’m going to see that possible victim’, as opposed to another sentence that says, ‘I saw that victim on the weekend’.
“Natural Language Processing will be able to say, the first is a pre-tense conversation, and the system will give that a much higher priority rating because it’s about an offence that is going to occur.
“This is using AI to risk score and assist with triaging what’s urgent – particularly in areas where agencies are having to deal with mass volumes of material. It allows agencies to deploy their finite resources appropriately to ensure the greatest level of public safety.”
From insight to prevention: Transforming policing models
John also highlighted the “speed to insight” that data and AI technology can offer: “We can go back to the former US Secretary of Defense Donald Rumsfeld’s statement of ‘unknown unknowns’ – something that some people ridiculed at the time, but have now come to understand as quite accurate.
“What AI, appropriate data management, and a properly regulated platform can give law enforcement is the ability to rapidly discover the unknown unknowns – uncovering emerging trends, revealing hidden networks and relationships, and extracting critical intelligence signals, very, very quickly.
Policing is shifting from a reactive model to one that is prevention led, and ethically governed AI is central to maintaining public trust.”
John Kilburn, SAS
“Protection and prevention are the two most important words in my view. Policing is shifting from a reactive model to one that is prevention led. And ethically governed AI is central to maintaining public trust,” John continued.
“Police data alone tells an incomplete story, shaped as it is by reporting behaviour, legal thresholds, and organisational capacity. When responsibly integrated with data from health, education, housing, social care, youth justice, and probation, a far richer picture of vulnerability and risk begins to emerge, highlighting patterns such as repeat victimisation, untreated trauma, and systemic service gaps well before they escalate into crime.
“In this context, the legitimacy of AI-driven insights depends not only on technical accuracy, but on transparency, accountability, and fairness in how data is sourced, combined, and used.
“Strong governance frameworks ensure that these capabilities are applied proportionately, protect individual rights, and avoid reinforcing bias, enabling police and partner agencies to act earlier, more effectively, and with greater public trust.”
Building trust: Transparency, accountability and governance
However, these benefits and many others share a common theme, as far as John is concerned; in all of these examples, AI is not making decisions.
“That’s the critical component. AI can help you compile information, identify linkages, and carry out a multitude of tasks and automation. But you still need to go back and validate the information, make sure you can account for where everything’s come from, and be confident that the governance around the system is truly sound.”
Once forces have a clear understanding of what they want their use of AI to achieve, that sound governance is part of the culture of explainability and accountability that John believes is crucial for delivering the necessary confidence in decision-making – both for the law enforcement workforce, and for the communities they serve.
Part of the community confidence around the use of AI will come from knowing that it is not making decisions for officers; it’s a tool, just like a flashlight or a police radio.”
John Kilburn, SAS
“Once you’re adopting systems and products that deliver specific benefits and employ AI, the next question you will be asked is how well is it governed? What is the accountability, the transparency, and the auditability around that product?
“Part of the community confidence around the use of AI will come from knowing that it is not making decisions for officers; it’s a tool, just like a flashlight or a police radio.
While John emphasises that it’s an important responsibility of agencies to ensure they implement appropriate systems with appropriate architecture, built in partnership with professionals who actually understand the technology, he is also clear that suppliers have a vital role to play.
“In our work with law enforcement agencies we’ve always understood that ‘trustworthy AI’ – technology that is designed to promote capability, prevent harm, ensure safety and follow ethical practices – must incorporate accountability, transparency, and be open to scrutiny.”
‘Suppliers’ in this context refers to technology companies that provide AI systems, data platforms, and analytics tools – including partners that help police organisations build data infrastructure, develop AI models, or deploy decision-support systems.
“It’s this governance that must be designed into AI systems from the start, and be embedded throughout the data and AI life cycle, without relying on third-party add-ons or external governance tools. These are the guardrails that communities expect, and that officers and staff can rely on.”
AI is already in policing – banning it is no longer an option
Understandably, police leaders may regard those demands for transparency, accountability and oversight as challenging in themselves; but without meeting those challenges, the alternative – ignoring AI or banning its use within an agency – is already doomed to fail.
“There are policing agencies out there that have said, ‘we’re not ready for this, we haven’t developed policies, don’t use anything that incorporates AI’. But officers are already using it on their phones, they’re using ChatGPT for example, and it’s being done outside the framework under which agencies are operating,” said John.
The genie is out of the bottle. If you think that producing a policy document saying ‘don’t use AI’ is going to stop it, you really need to go and meet some of your officers.”
John Kilburn, SAS
“The genie is out of the bottle. And if you think that producing a policy document saying ‘don’t use AI’ is going to stop it, you really need to go and meet some of your officers.”
But John also believes that, with clearly defined uses that align with an agency’s aims and objectives, and the right governance and guardrails in place, the ethical and effective use of AI could also help police forces to attract more recruits to the job.
“We know that one of the biggest problems facing law enforcement agencies is the recruitment and retention of new staff. But younger police officers are often far more tech savvy, they are keen to utilise AI, and it will make the life of the frontline officer easier.”
Beyond efficiency: Reshaping the workforce
Despite this additional benefit, there’s no suggestion from John that AI is any sort of ‘silver bullet’ that can answer all of policing’s performance problems.
“We do sometimes see people talk about AI and technology as being the solution to a particular problem. I see it more as a vehicle to change the way that policing commanders strategically think about how to combat the issues they have.
“It’s been said before, but AI is the most significant force multiplier for law enforcement agencies – if it is deployed ethically, and with the appropriate guidelines in place for its use.”
SAS – from data to trusted decisions
SAS is a global leader in data and AI, bringing software and industry-specific solutions alongside decades of law enforcement experience to its work with more than 1,600 public sector agencies in 134 countries. SAS helps policing organisations to transform data into trusted decisions — from using data to detect emerging crime patterns to designing ethical, transparent AI systems that assist in investigations and crime prevention.
To learn more about how SAS supports the public sector, visit Click Here.
