A Data-Centric Criminal Justice System
Data-centricity is today’s buzz-word of business, one of those statements that we hear bandied around as the magic bullet to fix everything, words which can mean all things to all people. However, for all the hype, we are genuine believers in the potential of data to transform organisations, indeed we make it happen time and again. Deciphering what ‘data sitting at the heart of the criminal justice system’ could tangibly mean, we deconstructed the ambition into three critical questions:
- What is the scope of a data-centric Criminal Justice System (CJS)?
Ministry Of Justice (MOJ) jurisdiction easily gives us courts, tribunals prisons & probation, but what else? Central government? Law enforcement? Local authorities? Health? Academia?...
- What data are we talking about?
The CJS is considered data-rich, but what is the focus? Bulk data? Targeted data? Management information?...
- What transformations do we aim to achieve through data-centricity?
This list genuinely could be endless, and we certainly aren’t scratching the surface yet, but what are the priority outcomes that we seek? Improved offender management? Strategic trends identification? Data-led policy creation? Data-driven decision making? Efficiency savings? Improved law enforcement outcomes?...
These questions naturally direct us towards a tangle of further considerations: a number of which are technical; many are ethical, and some are truly philosophical. Particularly taking question 3, the statement could easily be rephrased to ask ‘what is the priority purpose of the CJS?’; is it to punish, to rehabilitate, or something else entirely? We’re unlikely to answer this question through an 800-word blog.
1. Determining Scope
If we consider the first point above, how to bound the scope, the simplest answer would be to start with the jurisdiction of the MoJ: courts, tribunals, prisons and probation services. From an interoperability perspective, this would be the logical place to focus immediate efforts given the common frameworks, governance and funding, and clear distinct jurisdictions, which bind these areas together. Defining the operational links can help define the data pathways.
We can clearly see the benefits from more effective and efficient data sharing as both offenders and victims move through the system, and so does Richard Price, Director General of Performance, Strategy and Analysis at the Ministry of Justice, who has championed a clear strategy for the MoJ to become a Truly Data Led Justice System. Indeed, initiatives such as the MoJ’s Justice Data Lab and 2025 Data Strategy are already paving the way to data-centricity in the CJS.
However, the scope of data-centricity efforts could, and arguably should, be widened if the CJS is to truly be transformed through data. Improved data sharing and combined analysis with partners in policing, health and local authorities would be the immediate examples to jump to mind, however this is another list which grows exponentially the more thought you give it. Improved sharing of police data with courts could lead to improved, and more efficient, justice outcomes. Combining data from a number of these sources could help to identify mental health risks. In the same way as we talk about personalised medicine being the future of healthcare, could courts aspire to a parallel of personalised rehabilitation interventions by harnessing our data?
2. Defining Data
The second question we raise, defining what we mean by data, is a challenge not unique to the context of the CJS. The data which we share and harness will be driven by where it can be drawn from, what data is actually captured, the quality of the data available and the purposes we aim to use it for.
In particular, it’s worth pausing to think about the ethical considerations of a data-centric CJS. In almost all public sector work involving data, there are the immediate concerns around proportionality and intrusivity. We rightly expect parts of our lives, for instance medical records, to be treated with a level of confidentiality and privacy. Our free and democratic values mandate the state not to monitor our every move. However, how do we evaluate these ethical risks against the ethics of not doing analysis, of not taking action. How do we sensibly and effectively weigh up the risk of not using data to spot a high-risk reoffender, or an innocent person being found guilty, compared to perceptions of overly intrusive analysis?
3. Deciding Priorities
This final question is by far the most complex; technical and logistical questions such as enabling access to data, building capability, improving quality, are all solvable given time and funding. What will be essential to successfully transitioning to data-centricity will be clear prioritisation and leadership around the target outcomes. With competing priorities and stretched budgets, there needs to be forums for the sector to come together and discuss what challenges are of greatest need, and which opportunities offer greatest reward. There needs to be the senior buy-in and funding to enable these ambitions to become reality. And most of all, there needs to be involvement and leadership from ministerial level to determine, and mandate, the priority issues which data should be focused at.
What could this tangibly mean? Take the 2021 Strategic Rape Review, which took a system-wide approach to understand the shockingly low prosecution rates for rape crime. A ministerially led, laser focus to harness the power of data here could see things like analytics in Law Enforcement to automate behavioral analysis linking stranger rapes, helping to improve arrest rates. It could also see AI developed to help alert probation services of patterns indicative of reoffending, in turn improving the academic understanding of what these patterns are and how they are changing. Perhaps most importantly, it could see transformed use of management information to better constantly monitor the whole system, helping officials to evidence issues and make preventative interventions, rather than requiring reviews when things have already gone wrong.
This is one example of many, but only through taking an outcome driven approach will data-centricity be a success. We must build trust as tangible successes are unlocked incrementally. We must invest in outcomes, not just data-centricity, whatever we define that to mean.
Authors
