AI for Justice – bringing data into the courtroom
Blog: Capgemini CTO Blog
Despite the growing availability of data on people´s experience of justice, institutions worldwide still have a way’s to go in order to achieve justice for all. As a recent report by the World Justice Project estimates, five billion people have unmet justice needs globally. This includes 1.5 billion people who are unable to obtain civil, administrative, or criminal justice because they face obstacles to resolving their everyday justice issues. In light of the Sustainable Development Goals (SDGs), the justice gap shows that there is still a long way to go for the vision of an open and socially inclusive world.
In this article, we observe how the rise of big data and artificial intelligence (AI) has the potential to enhance this quest for sustainable justice systems. According to a study from 2020, AI could have a positive impact on 58% of the targets in the SDG 16 “Peace, Justice and strong institutions” (United Nations, 2015). If deployed ethically, AI can contribute to justice systems by improving accessibility, facilitating communication, and assisting administration.
AI to court! Mapping AI in the context of justice
While AI’s potential in fields such as environment and health is already widely explored, policy makers have also started to look at what data has in store for jurisdictional institutions. For example, Lithuania uses AI creatively for citizen services to have a positive impact on their well-being. Another example is Germany, specifying in its AI strategy that decisions and the appraisal of information will continue to be conducted by the staff of law enforcement authorities to maintain accountability and responsibility. Apart from these aspects, efficiency is also one of the main drivers for AI in justice systems, for example in Estonia, to tackle the growing complexity of court cases from the local to the European Union level, Estonia already embraces AI as a key solution to predict results of processes and discover new patterns.
In light of this dynamic, Capgemini organized a virtual workshop in partnership with the International Telecommunications Union (ITU) on the role of AI in achieving the SDGs, providing a snapshot of how AI can contribute to more accountable, accessible institutions and justice systems. According to the session survey, 78% of the participants saw AI as an enabler for justice in this context.
Gregor Strojin, chair of the Ad hoc Committee on AI at the Council of Europe and speaker at the AI4Good event, pointed out that, “while we should strive for increased efficiency, we should also keep in mind the essential aspect of justice in the eyes of citizens: They rely on trust and confidence in the functioning of the system.” Thus the question we need to answer on the journey to improved justice systems is: How can we make institutions in juridical processes more accessible, accountable, and efficient for citizens?
The approach needs to be human centred: As a gamechanger for work in courts, administration, and police, AI should be implemented at the service of citizens. Keeping this in mind and along the four PublicGoesAI playgrounds Capgemini envisions for AI in public services, the technology´s benefits can be leveraged through a wide range of applications.
1. Intelligently automating jurisdictional processes
Some estimates suggest that 69% of the work done by legal assistants can be automated by existing technology, including AI. The same is true for between 16% and 21% of the work done by judges. Through natural language processing (NLP) for example, algorithms can get to work to review documents and contracts, validate, and find relevant ones. Not only does this reduce the amount of time required to process cases, it also decreases the cost of providing legal services. To bring these benefits to a US court, Capgemini upgraded their system to enable case creation and e-filing of documents, as well as e-reminders for hearings.
Automation in justice systems makes institution more efficient and responsive, which in turn benefits citizen in their legal journey. Moreover, it gives the public servant more time to concentrate on the most crucial cases, while the trivial ones are carried out by AI.
2. Interacting with citizens to provide juridical guidance
When confronted with juridical processes, citizens often have difficulties navigating the legal system. This is confirmed by a recent study in the UK, finding that 70% of consumers would prefer using an automated online system to handle legal affairs instead of a human lawyer because of three important factors: speed, cost, and ease of use.
Such online systems include AI-powered chatbots, or even comprehensive AI guides, walking humans through critical decisions that need to be made for their particular circumstances. In this context, AI can be used to make juridical processes more explainable and accessible to citizens.
3. Detecting anomalies to prevent crime
Data is crucial to investigate complex matters and detect criminal incidents in due time. But although institutions dispose of a wide range of data, it is often unstructured and stored across organizational siloes. With AI, the full amount of data across different institutions can be used to detect anomalies. NLP can help extracting relevant information from unstructured data, or predictive analytics can enable crime investigators to detect suspicious activity by identifying and monitoring trends. For example, Manhattan´s District Attorney increased the number of human trafficking investigations they carry out yearly from 30 to 300, since using an AI-driven analytics software.
To integrate unstructured data into a common analytical environment, Capgemini developed the investigative tool Haystack. It is a customizable solution that provides seamless analysis for identifying anomalies. By applying AI to Haystack, investigations can be enhanced beyond data consolidation. Such AI solutions can alert law enforcement authorities to potential fraud and other criminal behavior, making justice systems more transparent and responsive with data-driven insights.
4. Supporting jurisdictional decision making
Judges need to take sensible decisions based on insights they analyse themselves. This makes decision making lack in-depth context and more vulnerable to bias. AI can access more of the relevant information to predict the outcomes of legal disputes and proceedings, thus supporting decision making in court.
According to a study from 2011, an AI trained on 200 years of supreme court records is already better than many human experts at predicting court decisions. Capgemini supported a Western European Ministry of Justice with a similar project: an analysis tool that fully uses available data from five years of law cases, consisting of around 12.000 decisions in court. The tool provides judges with a direct estimation of the amount of compensation for citizens. Given the historical data, the AI tool enables support in decision making, as well as analyzing biases in jurisprudence.
With great technology comes great responsibility
Whether to better inform the citizen, better detect crime or accompany decision makers in their case handling, one key requirement remains: AI must come with ethical guidelines and must be implemented in a fair and transparent way. If not appropriately done, technology might in the contrary echo human biases, with a damageable impact on our societies and institutions.
Ethics by design started to be promoted on an international level: Initiatives such as the Global Meeting on AI for Law Enforcement, hosted by UNICRI and INTERPOL encourage sharing best practices and provide toolkits for law enforcement authorities. The European Ethical Charter on the Use of AI in Juridical Systems and their Environment by the European Commission for the Efficiency of Justice (CEPEJ) is providing similar requirements for the use of AI in justice. In our workshop, Gregor Strojin provided insights into the work of the Ad hoc Committee on AI (CAHAI) of the Council of Europe and how its members develop frameworks along the ethical charter´s five principles.
A well-working and accessible justice system is a prerequisite for sustainable development as it reinforces civic values and promotes social order. Justice systems that embrace intelligent technologies have the potential to accelerate the achievement of SDG 16 “Peace, justice, and strong institutions.” If the benefits of AI are leveraged, it can contribute to more insights for decision makers in the justice area, more accessibility for the justiciable, and more accountability for institutions as whole.
The application possibilities of AI are already explored and evaluated, as they are operating in a complex environment which requires sensible decision-making. To unleash the technology´s full potential, policy makers should consider the following aspects when deploying AI to improve justice systems:
- Promoting the transmission of strategies into legal and political frameworks
- Identifying use cases that augment the jurisdictional actors with an added value in regards to decision taken
- Designing and implementing ethical AI guidelines to make sur AI serves the benefit of citizens and society.