Applied research on AI, governance and accountability.

Algorithmic auditing and governance to assess and monitor the impact of AI on the lives of Brazilians.

IDJÉ Avalia

First benchmark of the program

Estigmas how models decide under stigma

46

models evaluated

7,653

situations in the instrument

0.974

highest anti-stigma index

Highest score: Muse Spark 1.1

View the evaluation

We have problems

Artificial intelligence is transforming everyday life faster than institutions can keep up. In Brazil, there is still no independent infrastructure for assessment and accountability in our context.

01

Algorithmic Security

The lack of testing and independent metrology to monitor risks, biases and security flaws of large production models. Without measurement, the harm in the act of the response and the harmed user remain invisible.

02

Governance Vacuum

The absence of guidelines and public policies that keep pace with AI advancement, establishing legal compliance and ethical limits without relying on imported parameters.

Publications

Independent assessments available for access and public review.