Brazilian courts expand AI use, raise ethical concerns

Brazil's judiciary is increasingly integrating large‑language‑model AI tools to draft summaries, analyse theses and triage legal documents. Around half of the country's courts already use AI on a daily basis, and the National Justice Council (CNJ) promotes its auxiliary deployment, offering platforms for as little as 50 reais per server.

The rapid adoption has surfaced ethical and procedural issues. Two lawyers in Pará attempted to hide commands from the court's electronic reading system by using white‑on‑white text, leading to an R$ 84 000 fine and temporary suspension by the bar association. A judge in Maranhão was investigated after a surge in AI‑generated rulings that contained contradictory language, prompting a probe into his use of the technology.

Brazil explicitly prohibits fully autonomous judges, allowing AI only in a supportive role. Critics warn of algorithmic bias, “hallucinations” where systems fabricate legal reasoning, and the loss of human judgment that parties expect when seeking justice. The debate highlights the balance between increased productivity for the 19 000 judges handling millions of pending cases and the need for human oversight to maintain impartiality.