Brazilian AI for National Security and Democracy
Brazilian AI for National Security and Democracy
Democratic Sovereignty 5.0 — National Security from the Body-Territory
Brazilian democracy needs its own artificial intelligence.
Not an AI to replace judges, prosecutors, public servants, military personnel, researchers, or citizens.
But a sovereign AI to help us perceive patterns that cross decades, governments, companies, courts, contracts, social networks, financial flows, and major economic disputes.
Major attacks against a territory do not disappear simply because time has passed.
They leave traces.
They leave lawsuits.
They leave contracts.
They leave financial movements.
They leave networks of influence.
They leave suspicious decisions.
They leave incompatible enrichment.
They leave coordinated campaigns.
They leave links between companies, governments, law firms, public agents, media, social networks, and external interests.
Democratic Sovereignty 5.0 proposes that these traces be treated as historical materiality of the Body-Territory.
Not to automatically condemn anyone.
Not to create political persecution.
Not to violate rights.
But to prevent major structures of State capture from being erased by limitation periods, legal costs, defense asymmetry, or fragmentation of lawsuits.
Jiwasa: we perceive together
The lens of this blog is Jiwasa.
Jiwasa is the collective perception that no Body-Territory perceives everything alone.
One citizen sees one part.
One public servant sees another.
One judge sees part of the process.
One prosecutor sees another.
COAF sees financial signals.
The Federal Revenue Service sees tax patterns.
Social networks show reputation campaigns.
Notary offices show asset transfers.
Courts show repeated disputes.
Universities show methods.
Journalists show connections.
Communities feel consequences.
When everything remains separate, the attack appears small.
When we integrate signals with legality, the pattern appears.
This is the role of a Brazilian AI for national security and democracy: to help the Democratic Rule of Law see what fragmentation hides.
Lawsuits above one million as strategic materiality
Every lawsuit with estimated legal costs above the equivalent of one million reais should be treated as strategic materiality for sovereignty analysis.
This does not mean presuming guilt.
It means recognizing that lawsuits of this scale indicate disputes with high economic capacity.
Where there is a million-real legal dispute, there is asymmetry.
There is access to expensive law firms.
There is capacity to prolong litigation.
There is potential pressure on institutions.
There is risk of asset shielding.
There is a chance that the judicial system may be used as an instrument of power.
The proposal is that these lawsuits form a public, auditable, protected database capable of feeding a sovereign AI with legal metadata, without violating improper confidentiality.
The AI could cross-reference:
procedural history;
amounts involved;
corporate networks;
public contracts;
suspicious financial movements reported to COAF;
links with public agents;
repeated judicial decisions;
appeal patterns;
territorial connections;
reputation campaigns on social networks;
asset changes;
beneficial ownership;
relations with strategic sectors.
The goal is not to conduct a “fishing expedition.”
The goal is to build materiality through objective, public, proportional, and previously defined criteria.
It is not random investigation
The expression “fishing expedition” is used when someone searches without concrete grounds, hoping to find some indication by chance.
This proposal is different.
The concrete basis would be objective:
high economic impact lawsuits;
values above a threshold defined by law;
connection with strategic sectors;
risk to sovereignty;
relationship with public contracts;
financial indicators reported by competent bodies;
repeated patterns of action;
relevant territorial impact.
The AI would not search for culprits.
It would analyze already existing materialities.
Lawsuits already filed.
Contracts already signed.
Financial communications already made.
Public data already available.
Institutional records already documented.
The difference is that, instead of leaving everything scattered, the Brazilian State would begin to perceive patterns.
Legal prescription and State memory
Crimes may become time-barred under the law.
But the strategic memory of the State should not expire.
When a major attack affects the Body-Territory, its effects may last for generations.
A poorly conducted privatization.
An overpriced public work.
Industrial sabotage.
Currency manipulation.
Regulatory capture.
A bought judicial decision.
A harmful contract.
A coordinated campaign to destroy a strategic company.
Even when criminal punishment is no longer possible, the fact may remain relevant to national security, public policy design, prevention of new attacks, and protection of strategic sectors.
Brazilian AI would not serve to punish prescribed facts.
It would serve to prevent the State’s intelligence from forgetting patterns of capture.
Prescription may limit punishment.
But it does not need to erase institutional learning.
Sovereign, auditable, constitutional AI
An AI for national security must be Brazilian, auditable, and subordinated to the Constitution.
It cannot be an imported black box.
It cannot depend entirely on foreign clouds.
It cannot operate without public oversight.
It cannot produce automatic decisions without human review.
It cannot transform suspicion into guilt.
It cannot persecute political opponents.
It cannot violate fundamental rights.
Its role must be to:
detect patterns;
organize materialities;
generate alerts;
support public decisions;
protect data;
reduce technological dependency;
strengthen democratic institutions;
preserve the strategic memory of the State.
The final decision remains human, institutional, and constitutional.
AI proposes hypotheses.
The Public Prosecutor’s Office evaluates.
The Judiciary decides.
The defense contests.
Oversight bodies audit.
Society supervises.
Corruption as an attack on the Body-Territory
When governments, judges, prosecutors, businesspeople, or public agents are bought, the damage is not only moral.
It is territorial.
Corruption diverts energy from the Body-Territory.
It takes money from schools.
Weakens hospitals.
Delays public works.
Destroys industry.
Captures public policy.
Produces death, hunger, unemployment, dependency, and loss of sovereignty.
For this reason, major corruption schemes must be treated as threats to democratic security.
Not to militarize Justice.
But to recognize that systemic corruption can kidnap the State.
And when the State is kidnapped, we lose the capacity to care for what is common.
COAF, social networks, and contemporary materiality
COAF already plays a central role in producing financial intelligence against money laundering, terrorist financing, and related crimes.
But contemporary economic life also leaves signals in social networks, digital platforms, reputation campaigns, public ties, events, travel, visible wealth, contracts, and influence networks.
A sovereign AI could help identify convergence among:
financial signals;
public exposure;
corporate relationships;
judicial decisions;
State contracts;
digital campaigns;
asset flows;
participation in strategic sectors.
Always with legal limits.
Always with data protection.
Always with auditing.
Always with human review.
Democracy does not need to choose between naivety and abuse.
It can choose constitutional intelligence.
Reducing foreign technological dependency
Brazil’s Artificial Intelligence Plan 2024–2028 already points to the need for national infrastructure, supercomputing, algorithm development, and AI solutions for public services.
This is fundamental.
A national security AI cannot depend on foreign models trained with external values, interests, and priorities.
We need AI trained with Brazilian Portuguese, Latin American reality, national legislation, institutional history, protected public data, territorial diversity, and democratic commitment.
Technological sovereignty is political sovereignty.
Whoever controls AI infrastructure can influence what the State perceives, what it ignores, and how it decides.
Conclusion
Brazilian AI for National Security and Democracy must begin from a simple idea:
the State is us, organized to protect the Body-Territory.
We need AI that helps us perceive long-term attacks, structural corruption, institutional capture, economic assaults, influence networks, and threats to sovereignty.
Lawsuits above one million reais may form an objective layer of strategic materiality.
Even when old facts can no longer generate criminal punishment, they may continue serving as institutional memory to prevent new forms of capture.
AI does not condemn.
AI does not replace Justice.
AI does not conduct political persecution.
AI helps us see patterns that were previously hidden by fragmentation.
That is Jiwasa.
We perceive together.
We protect together.
We use science, technology, the Constitution, and memory to prevent the Brazilian Body-Territory from being kidnapped again.
References
Brazil. Constitution of the Federative Republic of Brazil of 1988.
Basis of the Democratic Rule of Law, popular sovereignty, due process, full defense, and institutional organization.Brazil. General Data Protection Law — LGPD, Law No. 13,709/2018.
Article 20 establishes the right to review decisions made solely on the basis of automated processing of personal data.Brazilian Artificial Intelligence Plan 2024–2028.
Provides for AI solutions to improve public services, social inclusion, high-performance infrastructure, and advanced algorithm development in Brazil.COAF. Integrated Management Report 2024.
Defines COAF’s mission as producing financial intelligence and supervising economic sectors to protect society against money laundering, terrorist financing, and related threats.National Council of Justice — Justice 4.0 Program.
Uses technology and artificial intelligence to bring the Brazilian judiciary closer to society and improve institutional efficiency.National Council of Justice. Artificial Intelligence in the Judiciary 2024.
Maps AI projects developed and used in the Brazilian Judiciary, indicating institutional progress and the need for governance.FGV Justiça. Artificial Intelligence in the Judiciary — 4th edition, 2025.
Analyzes AI tools in the Judiciary, including systems for triage, management, classification, and procedural practices.Data Privacy Brasil. AI with Rights, 2025.
Project focused on responsible and transparent use of AI in the Judiciary and protection of fundamental rights.Participa + Brasil. Public Consultation on Artificial Intelligence and Review of Automated Decisions, 2024.
Public debate on automated decisions, the LGPD, and impacts of AI on data subjects’ rights.FAPESC. Public Call No. 60/2025 — Program to Stimulate Technologies of Interest for National Sovereignty and Defense.
Recent reference connecting science, technology, AI, sovereignty, and national defense.SoberanIA — Government of Piauí.
Brazilian example of an AI model trained with Brazilian data and Portuguese language, seeking to respect the country’s reality and values.
