Brazil has the most fully worked AI-governance debate in its region and, as yet, no general AI law. For a country whose general AI law is still a bill, that international footprint and a real domestic case docket are the profile's centre of gravity.
The shape is an emerging jurisdiction whose general AI law is still a bill. The binding tier is mostly general law an AI dispute would meet, with two AI-specific rules that bind only within their domains, elections and the courts. The soft tier carries a well-funded national plan, and the draft tier carries the framework bill that would, once enacted, move much of this into a single statute.
Brazil's data-protection statute and the general law most likely to reach an AI harm. Its article 20 gives a right to request review of decisions taken solely on automated processing, with an explanation of the criteria used; the original guarantee of review by a natural person was removed in 2019, so the statute does not require a human reviewer.
Source: PlanaltoThe internet framework governing intermediary liability. In June 2025 the Supreme Court held article 19 partly unconstitutional and moved platform liability toward a notice-and-action regime. The fixed thesis addresses user content and platform duties and carries no AI-specific holding, so it reaches AI-amplified harm only indirectly.
Source: Planalto; STFThe one binding AI-specific transparency duty in force, scoped to electoral propaganda. Article 9-B requires explicit, prominent labeling of AI-generated or AI-altered campaign content; article 9-C bans campaign deepfakes, defined as synthetic audio or video altering the image or voice of a real, deceased or fictitious person, on pain of loss of registration or mandate. Administered by the electoral court, not a general operator duty.
Source: TSEThe binding rule for AI use inside the Judiciary. It classifies judicial AI by risk, requires meaningful human oversight with no decision taken solely by a machine, restricts sending personal data to external models, and creates a national AI committee. It binds the courts, not private operators.
Source: CNJThe authorship framework. Article 11 defines the author as the natural person who creates the work, and the law carries no computer-generated-works provision. Its limitations in article 46 are an enumerated list read restrictively, with no text-and-data-mining exception.
Source: Planalto; WIPO LexThe patent, trademark and design framework. Inventorship is confined to natural persons, and the INPI refused the DABUS filing that named an AI as inventor, withdrawing the national-phase application in 2022. No AI-inventorship provision.
Source: Planalto; WIPO LexThe first national AI strategy, built on the OECD AI Principles, setting direction across research, skills, public services and ethics. A strategy document with no operator duties.
Source: MCTIThe flagship AI investment plan, projecting around twenty-three billion reais over four years across five axes: computing infrastructure including a national supercomputer, diffusion and training, AI in public services, AI for business innovation, and support for the regulatory and governance process. An investment plan, not binding law.
Source: MCTIThe data-protection authority's AI work: a technical note analysing the AI bill, a study of generative-AI privacy risks under the LGPD, and a regulatory sandbox on AI and data protection with selected participants. Administrative and advisory, not a binding AI regulation; the authority's prospective role as AI regulator depends on the pending bill.
Source: ANPDThe flagship AI law. A risk-based framework with excessive-risk prohibited practices, high-risk duties, affected-person rights to information, explanation and contestation, an oversight authority role, sanctions of up to fifty million reais per infraction, and copyright provisions including a research and education text-and-data-mining lane. Not in force; a Chamber vote reportedly planned for 2026 had not occurred on the record.
Source: CongressAn executive complementary bill creating a national system to develop, regulate and govern AI, positioning the data-protection authority as a residual authority under a central council. It was sent to cure a constitutional initiative defect in the Senate text over assigning powers to the authority, and is intended to travel with the framework bill. Not in force.
Source: government releasesDraft guidelines sorting AI-related inventions into AI models, AI-based inventions and AI-assisted inventions, with patentability turning on human inventive contribution and fully autonomous AI output excluded. Emerging office guidance, still in consultation rather than settled.
Source: INPINamed founding signatory, agreement signed in Shanghai 16 Jul 2026; Brazil named among the twenty-nine founders in official releases, with domestic accession subject to internal procedures including ratification. Official source: Chinese government and Xinhua releases; Brazilian inter-ministerial press note.
Signatory, 1 to 2 Nov 2023. Official source: gov.uk, the Bletchley signatory list.
Signatory, 11 Feb 2025. Official source: elysee.fr, the statement signatory list.
Adherent; the national AI strategy is built on the principles. Official source: oecd.ai adherents.
Adopted 2021, as a UNESCO member. Official source: unesco.org.
Global Digital Compact adopted by consensus 22 Sep 2024; resolution 78/311 co-sponsored; resolution 78/265 adopted by consensus. Official source: UN records.
Adopted 18 Nov 2024 under Brazil's G20 presidency; reaffirms the G20 AI Principles and the UNESCO Recommendation and sets information-integrity and workplace-AI commitments. Official source: g20.org; the Rio declaration.
Adopted 7 Jul 2025 under Brazil's BRICS presidency; the first standalone BRICS AI-governance statement, calling for a UN-led framework and Global-South inclusion. Official source: brics.br; the statement text.
Axes: specificity (x, 0-100) is how squarely the instrument is written for AI; force (y, 0-100) is how hard it binds. Placements are editorial judgments for the comparison plot; the tier column is the authoritative classification. The pattern is a country whose AI-specific instruments cluster either high on force but narrow in domain (the electoral and judicial resolutions) or high on specificity but low on force (the bill, the plan, the office guidance), with a binding tier of general law beneath.
Tap a marker
The binding and prescriptive corner is the crowded one: the CAC-led measures put force and detail together, service by service. The sector guidance sits just left of the binding line, nonbinding in form but treated as obligatory. Dashed markers are drafts; the National AI Law would consolidate the whole stack into one statute.
Unlike a sparse jurisdiction, Brazil has a real and growing docket where artificial intelligence is materially at issue. Four adjudicated matters are recorded below, two of them with published process numbers, alongside a reported cluster. Per the fabrication screen, a process number is cited only where it is published and verifiable, and where a decided matter's number could not be confirmed it is described without one.
AI-enabled deepfake fraud is reported in the press as a growing consumer risk, charged where at all under the general fraud provisions; these are reported incidents, not adjudicated AI cases.
Brazil holds the human line in statute while its institutions move: authorship and inventorship stay with natural persons and the training exception waits in a bill, but the patent office already has AI examination guidance out for consultation.
The AI-IP Index reads a country on two questions that usually get folded together. Accommodation asks how far the law will go to protect intellectual property that a machine helped produce. Institution asks how much administrative machinery sits behind that law: its guidance, its international submissions, its disclosure duties.
A country can be generous on paper and have built almost nothing, or cautious and highly organised, so the two are scored apart and reported as a pair. Each rests on three questions, scored 0, 0.5 or 1.
How far the law will protect intellectual property that a machine helped make.
How much administrative machinery sits behind that law.
The scale runs the same way on both axes: 0 is the open or developed end, 1 is the restrictive or absent one, and 0.5 sits between for anything conditional or untested. Read as a pair, the two scores show whether AI-assisted work can be protected here, and whether the office has said so in writing.
Within each axis the three questions carry equal weight. That is a deliberate choice, not an oversight. An AI developer might reasonably value freedom to train above patent inventorship, but weighting is where indices lose their credibility, so the score stays a flat mean and the choice is stated in the open. Each cell also cites a statute section, a case or an office page, so any single score can be challenged without disputing the rest.
Two positions are worth naming. A country can be open on the law but thin on machinery, an open-but-undeveloped stance that bets on ambiguity. Or it can be restrictive on rights yet run an active office, a restrictive-but-developed stance that reflects a settled policy choice rather than a gap. The pair of scores tells those two apart where a single ranking would blur them.
A rubric is only as good as the results it produces on cases where the answer is already known. These three jurisdictions sit outside this series and are scored with the same six questions, as a check. Their law is the best documented anywhere, so if the scale puts them where the law actually sits, the scale is doing its job. They are reference points, not editions.
Firm judicial settlement at both poles and the weakest statutory position on training. The courts have closed the door on AI inventors and AI authors, while everything on training rides on unresolved fair-use litigation. There is no per se duty to declare AI use in a filing, only a candour duty that bites where the use is material.
Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022); Thaler v. Perlmutter (D.C. Cir. 2025), cert. denied 2 Mar 2026; Bartz v. Anthropic (N.D. Cal. 2025); USPTO revised inventorship guidance, 90 Fed. Reg. 54636 (28 Nov 2025); 89 Fed. Reg. 58128 (17 Jul 2024); 89 Fed. Reg. 25609 (11 Apr 2024).
The most permissive statutory position on training and the strongest transparency duty in force, sitting over an authorship question no court has answered. Read the label carefully: training and transparency run through EU law, but patents run through the EPO under the European Patent Convention, which is not an EU institution and covers 39 states.
Directive (EU) 2019/790, arts. 3 and 4; Regulation (EU) 2024/1689, art. 53(1)(c) and (d); EPO J 8/20 (2021); EPO Guidelines G-II 3.3.1 (2026 edition); Infopaq C-5/08; Painer C-145/10.
Liberal on patents since the Supreme Court abandoned the Aerotel test in February 2026, conservative on copyright, and currently thin on AI-specific machinery after the dedicated AI examination guidance was withdrawn. Section 9(3) nominally protects a work with no human author for 50 years, the widest AI-output right on paper, but no court has ever applied it to generative AI.
CDPA 1988 ss. 9(3), 12(7), 29A, 178; Thaler v Comptroller-General [2023] UKSC 49; Comptroller-General v Emotional Perception AI [2026] UKSC 3; UKIPO practice notice, 14 Jul 2026; Data (Use and Access) Act 2025 ss. 135 to 137.
The check earned its keep twice. The untested rule held: section 9(3) of the UK Copyright Act reads like the widest AI-output right anywhere, yet because no court has applied it to generative AI it scores as untested rather than open, which is the correct answer and the one a headline reading would miss. The check also exposed a limit worth stating. The Institution axis measures AI-specific machinery, not general office capacity, so a long-established office can score poorly when it has withdrawn its AI guidance and imposes no disclosure duty. Read that axis as what an office has built for AI, nothing wider.
Each row below carries the test we applied and the finding behind the score, with a source. The rationale, not the number, is the point.
Scores verified 21 July 2026, against AIP Index codebook v1.0. A score is only true as of its date.
What we look forA granted patent or a clearly workable examination route, not merely a silent statute.
The findingThe industrial-property law confines inventorship to natural persons, so an AI-assisted invention with a human inventor is open in principle while the machine cannot be named. The office refused the DABUS filing that named an AI as inventor, and its draft guidelines would admit AI-assisted inventions on a human-contribution test while excluding fully autonomous output. Open in principle, untested for a grant, and the guidance not yet settled.
What we look forProtection applied where human choices shape the work, and refusal of purely machine output.
The findingThe copyright law defines the author as the natural person who creates the work, so a human who directs an AI tool and contributes authorship can hold copyright, while purely AI-generated output has no qualifying author and no route to protection. There is no computer-generated-works provision and no ruling tests the point, so it is untested.
What we look forA statutory or judicial mining lane. Silence scores against the miner, since the reserved-rights default governs.
The findingThere is no training or text-and-data-mining exception in force. The copyright limitations in article 46 are an enumerated list read restrictively, with no mining carve-out; a research and education mining lane exists only in the draft framework bill. With copyright silent, the reserved-rights default governs the use of works for training.
What we look forA traceable, substantive submission on the record, not bare membership.
The findingBrazil and its industrial-property office are active internationally, presenting at the WIPO Conversation and leading a BRICS project on AI in patent examination, but no formal national written submission to the WIPO Conversation was found. Substantive participation without a submission on the record earns the middle.
What we look forA standalone AI examination text, not AI handled quietly under general practice.
The findingThe industrial-property office has published draft AI patent-examination guidelines for public consultation, sorting AI-related inventions by human contribution. This is real institutional action aimed at the AI-and-IP question, but it sits in consultation rather than as a settled directive, so it earns the middle rather than the absent pole.
What we look forA mandatory declaration inside the filing, not a proposal or an informal request.
The findingNo filing requires declaring AI use, and no in-force instrument imposes a labeling or transparency duty on AI-generated content in the intellectual-property system. The transparency duties that would reach developers exist only in the draft framework bill, and the in-force AI-labeling duty is scoped to electoral propaganda, not IP filings.
Method. Six questions, three per axis, scored 0, 0.5 or 1 against AIP Index codebook v1.0. The three questions in each axis are weighted equally by choice, and the two axes are reported as a pair, never blended into one number. Every score cites a statute, case or office page, carries a verified date, and records its direction of travel apart from the level. Positions are editorial judgements read from primary law and office practice, not official scores. WIPO's public catalogue of IP-office AI initiatives does not score jurisdictions on this basis, so the readings here come straight from the source law. Not legal advice.
Brazil has the most fully worked AI-governance debate in its region and, as yet, no general AI law. The flagship instrument, the Marco Legal da Inteligencia Artificial (PL 2338/2023), a risk-based framework influenced by the European approach, passed the Federal Senate in December 2024 and sits in the Chamber of Deputies awaiting the rapporteur's report, so its prohibited practices, high-risk duties and transparency obligations are not in force. What binds instead is a dense layer of general and sector law: a data-protection statute with an automated-decision review right, a reshaped internet-liability regime, and two binding AI-specific rules that reach only their own domains, an electoral resolution that labels AI content and bans deepfakes in campaigns, and a judicial-council resolution governing AI use inside the courts. The intellectual-property statutes confine authorship and inventorship to human makers. Brazil's distinctive weight is international: back-to-back G20 and BRICS presidencies with AI on the agenda, signature of the Bletchley and Paris texts, OECD adherence, and a named founding seat in the World AI Cooperation Organization. For a country whose general AI law is still a bill, that international footprint and a real domestic case docket are the profile's centre of gravity.
The site does three things: it maps every instrument that governs AI in Brazil, 20 of them, ordered by how hard each one binds; it reads the same instruments a second time by how squarely each is written for AI; and it records the matters where AI meets the law, with what was searched and found empty stated as plainly as what was found. Every claim traces to a primary source, and absences are recorded rather than papered over.
A membership claim enters this table only from an official list: a declaration annex, a treaty signature table, an organization's own record or an official government statement. Brazil's footprint is the largest in the series, anchored by consecutive G20 and BRICS presidencies with AI on the agenda.
Know of a ruling or proceeding in Brazil where AI is central? Send a source link; it will be reviewed before publishing.
Submit a case