Document redaction: A secure approach for UK police forces
We live in a heavily digitised society, with few exceptions left. As regulation moves to keep pace — for example, the arrival of GDPR and the updated Data Protection Act — the reality of digitisation is felt in the daily operations of some of our most critical public services.
Policing in particular has faced a notable step change. UK forces continue to grapple with the interwoven and daunting tasks of automating document processing and redacting millions of documents annually. These documents range from system-generated reports to handwritten statements.
The scale of the challenge can be difficult to comprehend. But take this one example; in 2022 there were over a million violent crimes recorded in England and Wales, according to the ONS. If each of those crimes generated an incident report, a crime report, a notebook entry and one statement, there are four million documents to process.
But police investigations are lengthy and complex. The MG series of forms, used to build case files for court, contains around 20 templates. It ranges from witness statements, to cover sheets, to disclosure schedules, which can easily run to hundreds of pages – or many more (famously, over 30,000 statements were taken in the Yorkshire Ripper case alone, which required the floor underneath the files to be reinforced).
Police forces log an extraordinary volume of victim crimes annually and the scale of managing these documents is challenging.
This process is not just time-consuming but fraught with the risk of exposing Personally Identifiable Information (PII) when documents are shared across agencies or the justice system.
It has been a known issue for the police for several years and is still seen as the number one digital problem by senior officers I speak with across the country.
Document redaction — a blue light response
Our perspective at Zaizi is vendor agnostic but we are also keenly interested in how available technologies can be practically put to work to: provide substantial resource savings; achieve shorter and more predictable returns on investment for public sector clients; and minimise PII exposure risks.
Our focus is on how such transformation work can be delivered following Secure By Design principles and to meet Police Digital Service’s rigorous assurance standards.
Given the volume of documentation handled by police, it’s easy to understand why manual redaction would be slow and prone to human error. Mistakes have significant implications for individuals’ privacy, safety and, more broadly, public trust.
A transformative approach using cloud infrastructure and pre-trained AI/ML tools (for example, the Oracle Cloud Infrastructure (OCI) Document Understanding offering) is capable of providing a more flexible blue light response to the automation of document redaction and management.
These technologies advance automation in file and document processing by classifying, summarising, and redacting sensitive content across various formats. This accelerates administrative tasks and improves accuracy, shifting human focus to quality assurance while cloud-based, sovereign deployments ensure scalability without performance loss.
The technology available now doesn’t just present the opportunity to streamline operations. It fosters a safer, more privacy-conscious handling of sensitive information and improves the consistency and traceability of the process too.
Transformation Day: Learn how emerging capabilities can serve police forces and the communities they protect
Secure by design: how police can protect data
Adopting new technologies invariably introduces new risks, particularly in cybersecurity. So, any approach must incorporate secure-by-design principles. This means embedding security at every level of the system, aligning with police digital service accreditation and internationally supported NIST guidelines.
Through PASF-approved supplier selection, application of secure landing zone patterns, robust encryption, access controls, and continuous security monitoring, it is possible to ensure that solutions for policing are as secure as they are efficient. But robust threat modelling is essential from the earliest design phases.
Even at a high level, the threat model for policing systems must consider various risks, including organised crime groups (OCGs), malicious and accidental insiders, supply chain risks, and even machine bias. To counter these, it is necessary to employ strategies such as rigorous access controls, detailed supply chain assessments, implementing secure development practices, formalising employee training, and deploying sophisticated monitoring systems.
A multi-layered, defence in depth approach ensures that systems not only detect and prevent external attacks but also mitigate risks from within the organisation.
AI opportunities for UK police forces
Automating document redaction with cloud-based AI and Machine Learning tools presents a significant opportunity for UK police forces. The adaptability of the technology offers a highly interoperable, evergreen, more efficient, safer, and secure solution.
Ongoing collaboration between police and digital specialists will be vital to delivering these innovations. For example, Transformation Day workshops can be a great starting point to see how emerging capabilities can serve police forces and the communities they protect.
As society continues rapid digitisation, critical public services like policing must modernise their approach.
Properly implemented, AI and cloud tools can benefit all stakeholders — reducing risks and resources for forces while upholding privacy and public trust.
We welcome conversations with the policing community on how to deliver efficient, impactive change as a core effort in our mission to make the UK the best and safest place to live and work.
Get in touch to find out more.
-
Agile government innovation: Tiny teams, mighty impact
-
Artificial intelligence: threats and opportunities in national security
-
The truth about government legacy IT systems — and what we’ve learned by working on them
-
AI and supply chain in the national security arena
-
Realising potential: Guiding local youth towards fulfilling futures
-
Digital government: Why liberating users and data from spreadsheets matters to me
