AI used in justice systems must stay transparent, accountable and under human control, Namibia’s Ministry of Justice and Labour Relations told an OECD gathering in Istanbul this week.
“Responsibility cannot be delegated to an algorithm. People should know when AI is used. They should be able to challenge an outcome. Higher risk requires stronger oversight,” said Audrin Mathe, the ministry’s executive director, at the 2026 OECD Global Roundtable on Equal Access to Justice.
Namibia is not an OECD member, but Mathe said AI’s spread through justice systems poses shared challenges that require common approaches and international cooperation.
Useful tools, new risks
Mathe named AI-powered case management, online dispute resolution, workflow automation and data analytics as tools that could make justice institutions more efficient and cut administrative burdens.
But he warned that introducing them without oversight creates new risks, particularly where automated systems influence legal processes or outcomes.
He also flagged a less discussed problem: as individuals and businesses use generative AI, courts and justice bodies could face a surge of inaccurate, overly complex or excessive submissions, overwhelming administrative systems and causing delays. He called for clear digital filing standards and proportionate verification of AI-generated legal content.
Measure outcomes, not pilots
Mathe urged justice institutions to stop running fragmented pilots and scale what works, through interoperable standards, secure infrastructure, staff training and continuous monitoring.
“We should measure timeliness, cost, accuracy, fairness and user experience. Success means better and more trusted outcomes,” he said.
Keep the human door open
He warned that digital reforms could shut out people without reliable internet, devices or digital skills.
“A digital doorway is useful only when a human doorway remains open,” Mathe said.
He pointed to Namibia’s State of Justice Summit, held from 16 to 18 September, as part of the country’s effort to examine how technology can improve justice without sacrificing accessibility or public confidence.
“Our priorities are clear. Scale what works. Govern AI responsibly. Keep human judgement at the centre. Protect access. The choice is not between innovation and rights. Innovation must serve rights. Efficiency must serve humanity.”
