Brazil's AI Legal Framework bill (PL 2338/2023) has had at least four scheduled vote dates since the end of 2025 — and none of them happened. As of August 2026, according to the Chamber of Deputies' own bill-tracking page, the legislation sits exactly where it's been for months: awaiting the rapporteur's formal report inside the Special Committee.
What changed
Nothing changed — and that's precisely the fact worth recording here. PL 2338/2023 was unanimously approved by the Senate in December 2024 and reached the Chamber of Deputies in March 2025. The Special Committee, chaired by deputy Luísa Canziani and reported by deputy Aguinaldo Ribeiro, held twelve public hearings between May and September 2025. Since then, the voting calendar has slipped at least four times: originally expected by the end of 2025, the vote was pushed to February 2026; then the rapporteur committed to delivering the report by May 19, with a vote on May 27; when that didn't happen, the new promise was June 9. None of those dates held, and the official tracking page on the Chamber's website still shows the bill stuck at the same stage: awaiting the rapporteur's report.
Why it matters
PL 2338 follows the EU AI Act model: it classifies AI systems by risk level (unacceptable, high, low/moderate), creates rights for people affected by AI decisions (transparency, explanation, contestation), establishes a National AI Regulation and Governance System (SIA), and sets fines of up to R$50 million per infraction. Meanwhile, the EU AI Act itself — the model that inspired Brazil's bill — already came into broad effect on August 2, as we covered here. The gap between "the law that inspired the bill is already in force" and "the Brazilian bill doesn't even have a formal report yet" is the most concrete measure of the delay.
The impact for Brazil
For anyone leading AI adoption at a Brazilian company, the practical lesson isn't to wait for the law before acting. Companies that delay AI governance investment until there's regulatory clarity — betting that "we'll adapt once the law is out" — are betting against a track record of four missed deadlines in eight months. And because PL 2338's text is explicitly modeled on the EU AI Act, much of the compliance groundwork — risk classification of the systems in use, technical documentation, contestation rights, audit trails — can start now, using the EU AI Act itself as a practical reference, without waiting for Brazil's final text.
Entercast's take
This is one of those situations where the absence of news is the news. This month alone we've covered here OpenAI's safety framework, Daybreak's access controls, and the auditable memory built into NVIDIA's NOOA — all examples of AI vendors building governance proactively, ahead of any legal requirement. The contrast with the progress of Brazil's AI Legal Framework is direct: the private sector, in Brazil and elsewhere, is defining its own governance practices at market speed, while the legal framework meant to give those practices predictability remains stuck with no exit date from the Special Committee. For anyone waiting for the law before acting: stop waiting.