The Internalisation and Externalisation of Judicial Functions in the Age of AI

This chapter examines how artificial intelligence is reshaping judicial functions through two interconnected trends: internalisation and externalisation. Internalisation refers to the integration of AI, especially generative AI and large language models, into courts and judicial decision-making. The authors argue that AI can improve efficiency, consistency, access to justice, and the management of judicial workloads. However, its use also raises serious constitutional concerns, including risks to due process, judicial independence, impartiality, transparency, privacy, and the continued role of human judgment. The chapter discusses examples such as the Colombian Constitutional Court’s ruling on judicial use of ChatGPT, as well as the EU AI Act’s classification of AI systems used in justice as high-risk. These examples show that AI may assist judges, but it should not replace human reasoning or responsibility. The authors emphasise the need for safeguards such as transparency, explainability, human oversight, data protection, and context-specific AI systems trained on reliable legal materials. The second major trend is externalisation, meaning the transfer of judicial or quasi-judicial functions outside traditional courts. This includes private platforms’ complaint-handling systems, out-of-court dispute resolution mechanisms, and the growing role of independent administrative authorities. While these mechanisms may provide faster and more accessible remedies, they also create constitutional risks. Private platforms may lack impartiality, transparency, and accountability, especially when automated systems are used to moderate content or resolve disputes. At the same time, administrative authorities under frameworks such as the GDPR, Digital Services Act, and AI Act increasingly become central actors in rights enforcement, potentially reducing courts to a secondary or residual role. Overall, the chapter argues that AI-driven justice must be assessed through constitutional principles. Efficiency alone cannot justify weakening the rule of law, fundamental rights, due process, or judicial accountability. The central challenge is to ensure that both AI tools inside courts and dispute-resolution mechanisms outside courts remain compatible with constitutional safeguards.