Helping Justice Agencies make the Jump into GenAI and beyond

Generative Artificial Intelligence (GenAI) solutions, specifically Chat-GPT, have dominated discussions across the technology sector and beyond in recent months, as organisations lean into conversations on what it might mean for them. The justice system should be no exception to those conversations. AI technologies are accelerating at phenomenal pace – offering huge potential within the public sector to automate repetitive tasks, streamline archaic processes and improve citizen interactions. Accenture research outlines that GenAI has the potential to disrupt ~75% of work activities in the public sector (A New Era of Generative AI for Everyone | Accenture). Yet key to the success of applying any form of AI solution in justice is doing it in a way that is responsible, secure, transparent and accountable so as not to erode public trust and confidence. Without trust, justice systems lose their legitimacy to uphold the rule of law and society loses confidence in the system’s ability to deliver fair justice.

Applications of AI within the Justice System

Possible applications of AI in justice range from simplifying and summarising legal language and jargon, helping people understand and navigate legal processes via chatbots, or addressing court backlogs by automatically prioritising court cases based on complexity, urgency and available resource.

When implemented successfully, AI can result in a number of benefits to support the delivery of justice:

• Empowered judicial employees – AI can tackle mundane, routine and time-consuming activities such as redacting sensitive data from court documents, freeing up court administrators to focus on high value tasks. 
• Improved experience for court users – AI can increase access to justice for people by simplifying complicated legal language, explaining court processes, or providing virtual assistance to fill out court documents.
•  End-to-end efficiencies – When used to automate time-consuming tasks, streamline processes or provide enterprise level command and control type analysis of capacity and demand, AI can improve judicial resource utilisation, resulting in significant cost savings which is critical in the context of public finances.

In October of this year our Global Public Safety Industry Practice hosted an event on GenAI in Justice at our new Data and AI Studio for Public Sector in Brussels. The event was attended by representatives from six European Justice organisations from across Europe including the UK, Belgium, Finland, Germany, Spain and EU-LISA. We provided demonstrations of AI solutions from across industries and facilitated a workshop to ideate possible use cases for AI in justice where we came up with ~30 ideas. Key themes from the session included using AI to improve citizen service delivery through personalised citizen engagement, analysis and prediction; enabling the justice workforce through productivity enhancements and decision-making support; enhancing operational efficiency through justice command centres and achieving mission objectives through bias detection and transparent performance analysis.

We are already moving from ideas to implementation – for example we have partnered with the Spanish Ministry of Justice to develop a GenAI-powered judicial search engine called Delfos. Using large language models and the Microsoft Cloud, Delfos gives users a quick, efficient way to access relevant information by searching hundreds of thousands of complex judicial documents via a simple and easy to use user interface.

Delfos takes a user’s question and quickly searches for and finds the relevant content in a matter of seconds. It then consumes this content—which is often jargon heavy—and rewrites it in plain language that is easy to understand. This has reduced the time it takes the workforce to locate specific information by 40%.

Steps towards valuable AI applications

So, what do justice agencies need to do to realise the value of AI technologies? Embarking on any AI transformation can be daunting, but based on our experiences, we have identified 3 steps that can help agencies implement AI in the most valuable way:

1. Prioritise use cases – justice agencies should start their AI journey by defining their strategic objectives and use them to assess and prioritise what use cases will best achieve them. Then, start small. Conduct ‘Test and Learn’ via low-risk projects with appropriate guardrails; with a focus on measuring impact and validating model selection and adaptation.

2. Rethink your workforce – the growing momentum of AI calls for a diverse, reconfigured workforce. GenAI will change the work people do. Equipping employees with the skills they need to use the technology and fostering a willingness to learn and seize longer-term transformational opportunities is essential.

3. Implement responsibly – justice agencies must design, develop and deploy AI ethically, securely and transparently to engender trust and scale AI with confidence. That means establishing the right governance structures early and making sure responsible principles are translated into development.

This is a pivotal moment for justice agencies to responsibly embrace the potential of GenAI and redesign how they operate to increase levels of performance, engagement, and satisfaction for all involved in the delivery of timely and fair justice.