James Slessor, Managing Director UK and Global Public Safety Practice, Accenture
The task of maintaining public safety has always been complex, today police leaders face a collection of challenges: a continually evolving threat picture, constrained finances, recruitment and retention pressure, an ageing digital infrastructure and a need to improve public trust and confidence. These challenges are compounded by the fact that vast amounts of data, of various types, is now an omnipresent feature of policing and society more broadly. Whether it’s evidence collection, bodycam footage, or the digital exhaust our devices create in our daily lives. Now more than ever technological innovation, and particularly technologies such as AI, present a significant opportunity to help police leaders address these challenges in ways that would have been unimaginable just a few years ago.
While the trade-off between innovation and maintaining current service provision may be challenging, embracing new technology is often one of the quickest ways to unlocking savings and efficiencies in resource-constrained times. Of course – doing so in a way that has robust responsible and ethical guardrails in place is critical to maintaining and building public trust and confidence. Evidence base policing approaches also have a critical role to play as well in understand, evaluating and promoting what is proven to work.
Police forces are already experimenting with innovative technology including AI, showcasing this in action. For instance, forces are trialling and testing new tools on the frontline, including: chatbots to manage non-emergency requests, automated translation, the analysis of crime hotspots and the use of facial recognition.
There are a growing number of future applications for AI, especially to help speed up investigations – for example supporting officers in performing procedural activities, as well as rapid analysis of digital evidence. The use of automated transcription and summarisation, support for statement taking, evidence collection, crime report/form completion, and recommendations for next best action could completely transform how investigations are undertaken today. The ability to transform victim support – for example, AI could provide victims with personalised referrals to victim support services based on their needs and location, guide victims through crime reporting processes and automate case management updates and connect them to support services. By offering a more personal, simple and faster experience, victims are far more likely to stay engaged until justice is done.
These examples of technical innovation and the use of AI can significantly reduce cognitive burden, put critical information in the hands of frontline officers to support data-driven decision making, and enable a greater focus on the public than administration. By adopting these advancements, policing can truly transform how it works, with an emphasis on bringing officers closer the communities and public that they serve.
How frontline policing harness the power of AI?
AI in policing invites both excitement and fear, with the technology’s potential to entrench potential biases, create a dependence on technology, erode critical thinking and create a more remote tech-automated response. But Policing is not alone in taking a cautious approach to the adoption of AI, however that pace of adoption is growing rapidly. Accenture’s research shows that globally, only 2% of organisations have operationalised responsible AI practices, but with a further 31% planning to do so by the end of 2024.
Based on over 700 global AI client engagements and our own research, we have identified 5 key imperatives that organisations must address to prepare for AI. Here is what these imperatives mean for policing and public safety agencies:
1. Responsible AI. AI use in policing raises ethical concerns, particularly around privacy, surveillance, and discrimination. Forces must take a human-centred approach and lean on diverse, multi-disciplinary teams to design, deploy, evaluate, and continuously improve solutions to minimise and expose bias. Rigorous guidelines such as the NPCC Covenant for Using Artificial Intelligence in Policing should be continuously adapted to keep up with technological change to inform every stage of AI development, from data collection to deployment.
2. Lead with Public Value. To maximise the benefit of AI, organisations must shift the focus from near-term focus on use cases to prioritising key strategic capabilities across the entire end-to-end service model, using public value, risk, transparency, and trust as the guiding principles to determine what solutions to leverage and invest in. This requires police forces to engage with communities early and often, ensuring AI initiatives align with public needs and expectations – including through active involvement of community representatives, industry, and third sector partners in the design and development of AI tools. Here to there is a critical role for Evidence Based Policing practices to really learn what works and build that into future deployments.
3. Establish a secure digital core. For UK policing, widespread challenges related to data quality, legacy systems, and data siloes represent major barriers to AI adoption and scaling. At a national level, forces should collaborate to develop a comprehensive strategy to consolidate and standardise data from diverse sources – incorporating a common set of data standards, access protocols and security markings – to foster seamless interoperability across forces to drive cost savings. Extended across the criminal justice system, such shared standards can enable more effective data sharing to paint a picture of end-to-end system performance and drive enhanced decision-making.
4. Reinvent talent. Our global research shows that despite 96% of workers saying they are ready to learn new skills to work with AI, only 5% of public and private sector organisations are actively reskilling their workforce at scale. Deepening understanding of AI across policing can not only help build trust and unlock innovation, but more broadly, supports the mission to build more digital and data-empowered police forces of the future. To prepare for an AI-enabled future, leaders and policymakers must plan to equip officers and staff with future-proof skills, ranging from prompt engineering and data analysis as well as the ability to work in and increasingly creative and hybrid manner with technology.
5. Drive continuous reinvention. Prioritising ideas and swiftly putting them into action -including through small-scale proof of concepts - is vital for forces to keep up with an ever-evolving technological landscape. These should only focus on safe low-risk use cases, and ultimately, AI should never totally replace human involvement due to the potential risks of using it within a public safety setting. Such experiments can demonstrate the value of AI to officers and staff with clear messaging from police leaders that using it will make their job easier and free them up to focus on citizen facing tasks in their communities. With the public, communicating transparently regarding the AI’s purpose and how the algorithm works (i.e. that it is not in a black box), and explaining the responsible AI safeguards that are in place will be vital to bringing the public with us. It is also important that these pilot when proven successful, can also be successfully scaled with greater benefits to all.
The role of evidence-based policing in helping to assess the value of innovative technology, the use of AI and fundamentally what works is critical. It is great to see this as a theme at the Society for Evidence Based Policing 2024 Conference in Cambridge where Accenture is excited to be participating. Crucially, such cross-sectoral forums present vital opportunities for diverse exchange and debate on the state of AI and technology innovation across UK policing – and how collectively, we might chart a clear path through the disruption to promote what works and deliver greater benefit to the communities and public that policing serves.