Prompt injection in the Judiciary - Risks of generative AI in procedural documents and impacts on legal information security.

The incorporation of artificial intelligence into the Judiciary broadens the debate on efficiency and innovation, but also highlights new challenges related to information security and digital governance.

Among them is prompt injection , a technique that uses hidden commands in documents to interfere with the behavior of AI systems. Brazilian courts have already identified instances involving this practice in electronic petitions, broadening the discussion about informational integrity and the responsible use of automated tools in the justice system.

We've prepared a carousel explaining how the technique works and the legal and technological risks associated with the use of AI in the Judiciary. Check it out in the PDF below:

Prompt Injection e Judiciario.pdf

Download PDF • 2.08MB