Eskwai for Students: RAG AI assistant uses 12K case laws for Ghana law students

A team of researchers led by George Boateng has developed Eskwai for Students, a generative AI assistant tailored for legal education in Ghana. Unlike general-purpose chatbots, Eskwai is a retrieval-augmented generation (RAG) system that grounds its answers in a curated database of over 12,000 case laws and 1,400 pieces of Ghanaian legislation. This ensures responses are legally accurate and contextually relevant. The team deployed the system in a longitudinal study spanning 30 months (2.5 years), during which 3,100 law students used the tool to submit 32,000 queries. The work, accepted at the 27th International Conference on AI in Education (AIED 2026), evaluates both the helpfulness of the AI and the ethical concerns raised by student queries, such as requests for direct legal advice rather than educational guidance.

The study provides one of the first large-scale insights into how law students in the Global South use generative AI for their studies. By focusing on Ghana's legal framework — rather than relying on general or Western-centric legal data — Eskwai for Students aims to bridge a critical gap. The researchers highlight that while the tool is effective for learning and research, it requires careful oversight to prevent misuse, such as students seeking answers to exam questions or relying on the AI for real-world legal counsel. This work contributes to the responsible deployment of AI in education, particularly in under-resourced legal systems, and offers a blueprint for similar initiatives across Africa and beyond.