
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
In Pará, Brazil, two lawyers were fined for inserting hidden commands in legal documents to manipulate the Galileu AI system used in judicial processes. This misuse led to sanctions and was later cited by a Connecticut judge in the US to address a similar AI manipulation attempt, highlighting international legal concerns over AI misuse in courts.[AI generated]
Why's our monitor labelling this an incident or hazard?
The event explicitly involves an AI system (Galileu) used by the judiciary to analyze legal petitions. The misuse of hidden AI commands in petitions to manipulate AI responses constitutes a direct misuse of AI technology. The detection and blocking of these commands by the AI system led to legal and professional consequences for the lawyers, which are harms under the framework (violation of legal and professional obligations). The involvement of AI is central and directly linked to the harm caused. The US judge's decision referencing the Brazilian case further confirms the significance of the AI system's role. Hence, this is an AI Incident.[AI generated]