Can artificial intelligence make ethical legal decisions, and what happens when it gets them wrong? These questions are becoming increasingly urgent as universities, policymakers, and legal experts grapple with the expanding role of AI in justice systems.
At the International MaxUp Legathon held in Astana, Kazakhstan, students from 13 countries examined whether AI could ever replace judges or lawyers in courtrooms. Hosted at Maqsut Narikbayev University, the event focused on how emerging technologies are reshaping legal systems, human rights, and ethical frameworks across different jurisdictions.
The idea of an AI-run courtroom—where algorithms serve as jurors, and automated systems act as lawyers—was central to discussions. While such scenarios may appear futuristic, participants stressed that the foundations of justice rely heavily on human reasoning, empathy, and moral judgement.
Experts at the event argued that AI lacks emotional understanding and cannot meaningfully assess mitigating circumstances. Instead, machine systems operate by analysing patterns in large datasets, producing outputs based on statistical correlations rather than moral reasoning. Concerns were raised that if training data contains errors or bias, AI systems could replicate and reinforce flawed outcomes.
Sergey Pen, Deputy Chairman at Maqsut Narikbayev University, said AI currently cannot replicate legal reasoning. He noted that while language models can process vast information quickly, they fail to construct the logical legal justification required in judicial decisions. According to him, AI should remain a supportive tool rather than an authority in court rulings.
Several countries are already experimenting with limited legal applications of AI. In Kazakhstan, systems are used to review case law and support judges by identifying similar precedents. In China, students reported that AI assists with administrative tasks such as sorting cases and retrieving relevant legal documents, but does not determine verdicts.
Despite these developments, students from Georgia and Canada highlighted ongoing concerns about whether AI could ever achieve legitimate ethical decision-making. They pointed to a core limitation: algorithms lack moral awareness, raising doubts about whether justice can ever be fully automated.
A major issue discussed was accountability. In traditional legal systems, judges are responsible for their rulings and can be held liable through appeals or disciplinary procedures. With AI systems, responsibility becomes unclear—whether it lies with developers, institutions, or users remains unresolved.
Some participants suggested that developers should bear primary responsibility for harm caused by AI systems, while others called for stronger legal frameworks to define liability in automated decision-making.
Kazakhstan’s recent legislation on artificial intelligence reflects this cautious approach, defining AI strictly as a tool that supports human decision-making rather than replacing it. The law emphasises human-centred control over algorithmic systems.
As AI continues to advance, the Legathon discussions underline a central concern: while technology may assist justice, the authority to judge must remain human.
