On May 22, 2026, Senator Professora Dorinha Seabra (União-TO) filed PL 1.897/2026, a bill that would amend Brazil's 1989 anti-discrimination statute (Lei 7.716) to criminalize profiting from hate speech online. The base penalty is two to five years in prison plus a fine. Crucially, liability does not stop at the author: it reaches anyone who intermediates, sponsors, finances, contracts the promotion of, or manages affiliate programs for discriminatory content — including content targeting gender or sexual orientation — and anyone who provides the underlying infrastructure. Penalties rise by one-third to one-half when paid ads, bot networks, inauthentic accounts, or obscured funding are involved.
The bill lands on top of an already-shifting regulatory stack. In June 2025 the Supreme Federal Court (STF) held Article 19 of the Marco Civil partially unconstitutional, and on May 20, 2026 the federal government issued Decrees 12.975 and 12.976 to operationalize that ruling. PL 1.897 is the third layer — and the first to attach prison time to the chain of intermediaries.
The case for following the money
The strongest argument for the bill is genuine. Coordinated hate campaigns are not spontaneous; they are often a business. Affiliate networks, ad placements, and promoted posts turn discrimination into recurring revenue, and the people who finance and distribute that content are frequently more culpable — and more solvent — than the anonymous accounts that post it. Targeting the economic core, as Seabra argues, aims at the part of the machine that actually scales harm. Brazil also has hard experience here: misogynistic "red pill" content and gender-based digital violence have produced real-world victims, and the existing civil regime has struggled to deter organized, monetized abuse.
That diagnosis is correct. The problem is the instrument.
A criminal statute built on an undefined term
Brazilian law does not contain a precise, settled definition of "hate speech." The STF's June 26, 2025 ruling itself folded hate crimes into a broader category of "serious" content — alongside anti-democratic acts, terrorism, incitement to self-harm, gender-based violence, child sexual abuse, and human trafficking — for which platforms can face liability without a court order when there is a "systemic failure" (Global Network Initiative analysis). Even sympathetic observers concede that "systemic failure" and "adequate measures" remain undefined, with no procedure for assessing them.
Attaching civil liability to a fuzzy standard is one thing; attaching prison to it is another. Criminal law is supposed to be the domain where notice and precision matter most. A statute that exposes "financiers," "affiliate-program managers," and "infrastructure" providers to two-to-five-year sentences for content whose legal boundary is contested asks intermediaries to predict, under threat of incarceration, how a future court will classify third-party speech. The rational response is not careful moderation — it is pre-emptive over-removal and de-banking of any account near the line.
The chilling math
The reach to infrastructure and payment intermediaries is where proportionality breaks down. A hosting provider, an ad exchange, or a payments processor cannot adjudicate whether a given post crosses from offensive-but-lawful into criminal discrimination. Faced with criminal exposure plus a one-third-to-one-half enhancement for any monetized distribution, these actors will route around risk by cutting off marginal, controversial, or simply ambiguous accounts wholesale. That collateral damage falls hardest on small Brazilian publishers, activists, and minority voices who depend on the same ad and payment rails — the very groups hate-speech law is meant to protect.
This is the recurring lesson of intermediary-liability design: the broader and vaguer the duty, the more lawful speech gets swept up in the compliance margin. The 2025 STF decision already pushed Brazil from a court-order model toward notice-and-takedown and presumed liability for paid promotion and bot-driven distribution. The May 2026 decrees layered on operational duties — permanent complaint channels, moderation transparency, and a one-year retention requirement for ad and advertiser records, supervised by the data-protection authority ANPD (Decree 12.975/2026). Each layer independently raises the cost of being wrong. PL 1.897 would convert that cost from money into liberty.
A more proportionate path
None of this argues for impunity. The monetization chain is a legitimate target, and Brazil is right to want enforcement that reaches financiers rather than only the lowest-paid poster. But the tools should match the harm. A better-targeted version of this bill would: (1) define the predicate offense by reference to specific, already-criminal categories rather than the open-ended "hate speech"; (2) require knowledge or willful blindness, not mere participation in an infrastructure chain, before criminal liability attaches; and (3) keep generic infrastructure and payment intermediaries in a civil, notice-based regime, reserving prison for those who knowingly orchestrate and profit from unlawful campaigns.
Civil society is not uniformly opposed to the broader project — the 40-plus organizations of the Coalizão Direitos na Rede backed the May decrees as implementing the STF ruling with due-process safeguards. That is the model to extend: clear procedures, defined terms, and proportionate sanctions. PL 1.897, as filed, inverts the priority — maximal criminal reach atop minimal definitional clarity. For a publication that believes regulation should be evidence-based and proportionate, the verdict is straightforward: keep the target, change the weapon.