Artificial Intelligence Development and Trust Act, 1757

Artificial Intelligence Development and Trust Act, 1757

13th Cortes Federales
Long title	An Act to support the development of artificial intelligence and the domestic artificial intelligence industry; to establish a foundation of public trust in artificial intelligence through duties of safety and transparency; to restrict harmful uses of artificial intelligence; to establish the Federal Council on Artificial Intelligence and the Federal Office for Artificial Intelligence Safety; among other purposes.
Introduced by	The Council of State
Extent	Entire Federation of Nouvelle Alexandrie
Dates
Royal Assent	1.IX.1757 AN
Commencement	1.IX.1757 AN
Other legislation
Amendments	None
Related legislation	Aldurian Economic Diversification and Recovery Act, 1755 Proclamation of Punta Santiago

The Artificial Intelligence Development and Trust Act, 1757 (Alexandrian: Loi sur le développement et la confiance de l'intelligence artificielle, 1757; Martino: Ley de Desarrollo y Confianza de la Inteligencia Artificial, 1757; Wechua: Kamasqa Yachay Kamachiy, 1757) is a law of Nouvelle Alexandrie that establishes the Federation's first general framework for the development, deployment, and oversight of artificial intelligence. The Act pairs measures to expand the domestic artificial intelligence industry with duties of safety and transparency on those who build and use such systems, and restricts a defined set of harmful uses. It creates two federal bodies, the Federal Council on Artificial Intelligence and the Federal Office for Artificial Intelligence Safety.

The Act was brought forward by the Council of State on 18.VIII.1757 AN, with Secretary of Research and Development Alexander Melas as its lead sponsor and the Department of Justice responsible for its enforcement provisions. The government presented the bill as the computing and software companion to the hardware diversification under way through the Aldurian Economic Diversification and Recovery Act, 1755, part of a wider effort to move the Federation's economy beyond raw Alexandrium exports after the market collapse of 1757 AN.

The bill was considered by the 13th Cortes Federales. Its Federal Assembly debate opened on 25.VIII.1757 AN and closed on 29.VIII.1757 AN, with the Speaker ordering final division on the amended text. The Federal Assembly passed the amended bill on 1.IX.1757 AN by 687 votes to 0, with 62 abstentions or absences recorded, and the Act received Royal Assent and came into force the same day. Its spending provisions, including a public co-investment fund and a national computing initiative, were drawn as a capped envelope released in itemized tranches, the structure the Civic Governance Alliance had required of the government's earlier spending bills.

Background

The Federation entered 1757 AN as a high-research economy with little law governing the digital technologies it had come to rely on. Research and development spending stood at 8.4 percent of gross domestic product, and Alexandrium-accelerated computing had spread into defense, medicine, intelligence analysis, and finance. None of it sat under a dedicated statute. A private cryptocurrency market worth tens of billions of New Alexandrian ecu operated with almost no oversight, and the spread of Alexandrium health monitors had opened a public argument over biometric data, insurance pricing, and corporate surveillance that reached the Federal Assembly in 1746 AN.

Government interest in artificial intelligence predated the bill. Premier Juan Pablo Jimenez had created the Royal Commission on Artificial Intelligence in 1745 AN to study how such systems might reduce the administrative burden of the public sector. The Commission's remit was advisory and confined to government efficiency. It produced no general rules for private developers, no safety standards, and no body to apply them.

The immediate occasion for the Act was the Alexandrium market collapse of 1757 AN, a price and demand failure that followed the withdrawal of two large Natopian buyers and tipped the Federation into a difficult year. The crash gave weight to a case New Alexandrian scientists had been making since 1751 AN, when the Institute for Strategic Studies projected that domestic Alexandrium extraction would peak around 1768 AN. The government's answer had so far been built in hardware. The Aldurian Economic Diversification and Recovery Act, 1755 funded the Aldurian foundries, Foundation 1768, and the Institute for Wide-Bandgap Electronics, and the Cárdenas Protocol of 14.II.1757 AN brought Natopian fabrication technology and co-investment into the same effort. That work, as its own directors stressed, concerned power electronics rather than computing. The Act was framed as the layer above the silicon and diamond: the software, services, and computing capacity through which the Federation might add value rather than ship raw material.

The bill drew on the Constancian, Vegnese, and wider Raspur Pact discussion of artificial intelligence law, but its restrictions were shaped by domestic controversy. The Proclamation of Punta Santiago guarantees a Right to Privacy and Secrecy of Correspondence, and the health-monitor debate had already turned those guarantees into a live political question. The New Alexandrian Civil Liberties Union and a number of deputies had pressed for limits on biometric tracking and on the use of health data in insurance pricing well before the bill was drafted.

Key provisions

The Act applies to artificial intelligence developed, supplied, or used within the Federation. It divides those it regulates into developers, who build artificial intelligence systems, and operators, who put such systems to use in products and services. It reaches beyond the Federation's borders in one respect: a foreign provider whose systems are used in the Federation above a threshold set by regulation must appoint a New Alexandrian representative answerable to the authorities.

The bill establishes two bodies. The Federal Council on Artificial Intelligence is a standing inter-departmental body chaired by the Vice-Premier, with the Secretaries of Research and Development, Justice, Trade and Industry, the Interior, and Defense among its members. The Council prepares a National Artificial Intelligence Strategy every three years, coordinates the work of the departments, and absorbs the functions of the Royal Commission on Artificial Intelligence, which the Act dissolves. The Federal Office for Artificial Intelligence Safety sits within the Department of Research and Development. The Office writes technical standards, runs a voluntary certification scheme, operates the regulatory sandbox, receives reports of serious incidents, and conducts the Federation's safety evaluation of advanced systems. It is a standards and certification body rather than a heavy enforcement agency; serious breaches are referred to the Department of Justice.

The Act defines a category of high-impact artificial intelligence: systems used in fields where a failure or misuse could affect life, safety, or fundamental rights. The named fields are the supply of energy, healthcare and medical devices, financial services and credit, biometric identification, critical infrastructure, decisions by public bodies that affect a person's rights or access to services, and the screening of candidates in employment or education. Operators of high-impact systems must keep a risk-management process, maintain documentation, provide for human oversight, give affected persons a plain explanation of decisions that concern them, and carry out an impact assessment before deployment. The Office may add fields to the high-impact list by regulation, subject to confirmation by the Cortes Federales.

A short set of practices is prohibited outright. The Act bars artificial intelligence designed to manipulate behavior through means a person cannot perceive, social scoring of individuals by public bodies, and the creation of synthetic intimate images of a real person without consent. Real-time remote biometric identification of people in public spaces is prohibited except to the Federal Gendarmerie and the intelligence services acting under a warrant from a federal judge, on grounds specified in the Act. These restrictions are tied in the bill's recitals to the Right to Privacy under the Proclamation of Punta Santiago.

The bill responds directly to the health-monitor controversy. It bars the use of artificial intelligence inferences drawn from biometric or health data to set insurance terms, deny coverage, or screen for employment, and it requires consent and data minimization for systems that process such data. Generative systems carry a transparency duty: content produced or substantially altered by artificial intelligence must be marked as such, and synthetic media presented as a real person or event in an election is prohibited unless clearly labeled.

The development half of the Act assembles four measures. It establishes a regulatory sandbox, administered by the Office, in which a developer may test a system under supervision and on relaxed conditions for a fixed period, with a fast-track route to market for low-risk applications. It creates a National Compute and Data Initiative that pools existing Alexandrium-accelerated computing capacity for shared access by startups, universities, and public researchers, and sets up an open public-data framework under privacy safeguards. It opens a fast-track residency for foreign artificial intelligence specialists and funds a federal training and retraining pipeline through the Department of Education, Sports, and Culture and the universities, among them the Royal University of Parap. It provides a research and development tax credit for qualifying artificial intelligence work and establishes a public Artificial Intelligence Co-Investment Fund that invests alongside private capital in New Alexandrian firms. The Fund and the Initiative are capped, and their money is released in tranches authorized one at a time, each presented with its costs. After amendment in debate, supported projects must publish resource and productivity statements before approval, and the Office must report annually to the Cortes on compute allocation, energy demand, regional distribution, and measurable public or economic value.

Enforcement is staged. The Office leads with guidance and certification before penalties, and the strongest sanctions are reserved for the prohibited practices and for high-impact failures that cause harm. Administrative fines, orders to suspend or withdraw a system, and, for the gravest breaches, referral for prosecution are available. Decisions of the Office and the Department may be appealed to the courts. The governance and development provisions are timed to take effect before the duties on high-impact operators, giving industry a transition period; the prohibited practices take effect on Royal Assent.

Legislative progress

Legislative progress of the Artificial Intelligence Development and Trust Act, 1757

Stage	Date	Chamber	Action	Votes (Y-N-A)	Details
1	18.VIII.1757 AN	Federal Assembly	Introduction	-	Introduced by the Council of State
2	24.VIII.1757 AN	Federal Assembly	Committee referral	-	Referred to the Budget and Finances Committee for its fiscal provisions
3	25.VIII.1757 AN	Federal Assembly	Debate opened	-	Introduced on the floor by Alexander Melas, Secretary of Research and Development
4	27.VIII.1757 AN	Federal Assembly	Sponsor amendment received	-	Government amendment to Articles 17 and 19 on resource and productivity statements and annual reporting
5	29.VIII.1757 AN	Federal Assembly	Final division pending	Pending	Speaker ordered final division on the amended text

The bill reached the chamber in the middle of the Federation's hardest year in a decade, and its progress was bound up with the same coalition arithmetic that governed every spending measure of the 13th Cortes Federales. The government of Premier José Manuel Montero held its majority through the Aldurian Independent Coalition, whose coordinator Émile Bouchard had backed the diversification program on the understanding that it would produce working industry rather than promises. The development provisions of the Act, presented as the next stage of that program, drew the Coalition's support, with Bouchard pressing as before for tangible results and for benefits that reached beyond Cárdenas.

The floor debate opened with Melas presenting the bill as a framework for growth and trust: a Federal Council on Artificial Intelligence to coordinate policy, a Federal Office for Artificial Intelligence Safety to set standards, and duties on systems used in health, finance, employment, and other high-impact fields. The FCP, speaking through Tomás Cubero, supported the bill but pressed the government on audit capacity, regional access to compute and co-investment funds, synthetic-media labeling outside elections, sandbox liability, algorithmic bias outside health data, and social scoring by private companies.

Melas answered that the Government would accept language requiring the Office to publish an annual audit plan, regional reporting for the National Compute and Data Initiative and the Artificial Intelligence Co-Investment Fund, and rules extending suspicion of social scoring to private ranking for housing, credit, insurance, employment, education, and essential services. He also argued that Article 11 already covered generative content generally, that Article 14 added an election-specific rule, and that sandbox supervision did not confer immunity for harm after deployment.

The only sponsor amendment formally received before final division came after Vice-Premier Fred Strong warned that artificial intelligence policy could add an energy and resource burden to the real economy while rewarding systems that produced managerial reports rather than practical value. Melas tabled a Government amendment to Articles 17 and 19 requiring projects supported by the National Compute and Data Initiative or the Artificial Intelligence Co-Investment Fund to publish resource and productivity statements before approval. The Speaker received the amendment, and no objection was maintained before debate closed.

Bill text

Artificial Intelligence Development and Trust Act, 1757

THE READING OF
A
BILL
TO
support the development of artificial intelligence and the domestic artificial intelligence industry; to establish a foundation of public trust in artificial intelligence through duties of safety and transparency; to restrict harmful uses of artificial intelligence; to establish the Federal Council on Artificial Intelligence and the Federal Office for Artificial Intelligence Safety; among other purposes.

Artificial Intelligence Development and Trust Act, 1757

Presented by the Council of State
Ordered, by the Cortes Federales of Nouvelle Alexandrie,
to be Printed, 18.VIII.1757 AN.

BE IT ENACTED by the King's Most Excellent Majesty, by and with the advice and consent of the Cortes Federales, in this present session assembled, and by the authority of the same, as follows:-

PART I
GENERAL PROVISIONS

Article 1: Citation.

1. This Act may be cited as the "Artificial Intelligence Development and Trust Act, 1757".

Article 2: Definitions.

1. For the purposes of this Act:
   1. "artificial intelligence" means a machine-based system that, for explicit or implicit objectives, infers from the input it receives how to generate outputs such as predictions, content, recommendations, or decisions;
   2. "artificial intelligence system" means software, alone or as a component of a product, that operates using artificial intelligence;
   3. "developer" means a person who builds an artificial intelligence system or has one built to place it on the market under their own name;
   4. "operator" means a person who uses an artificial intelligence system in the course of a product, service, or activity, other than a person using it for purely personal purposes;
   5. "high-impact artificial intelligence" means an artificial intelligence system used in a field listed in Article 9;
   6. "generative artificial intelligence" means an artificial intelligence system that produces text, images, audio, video, or other content;
   7. "synthetic media" means content that has been produced or substantially altered by an artificial intelligence system so as to depict a real person or event;
   8. "remote biometric identification" means the identification of a person at a distance by comparing their biometric data against data held in a reference store, without their active participation;
   9. "the Council" means the Federal Council on Artificial Intelligence established under Article 5;
   10. "the Office" means the Federal Office for Artificial Intelligence Safety established under Article 7.

Article 3: Purpose.

1. The purposes of this Act are:
   1. to support the sound development of artificial intelligence and the growth of the domestic artificial intelligence industry;
   2. to establish public trust in artificial intelligence through duties of safety, transparency, and accountability;
   3. to protect the rights, safety, and dignity of persons in the Federation from harm caused by artificial intelligence;
   4. to give effect, in the field of artificial intelligence, to the Right to Privacy guaranteed by the Proclamation of Punta Santiago.

Article 4: Scope and application.

1. This Act applies to artificial intelligence developed, supplied, or used within the Federation.
2. This Act does not apply to an artificial intelligence system developed and used solely for personal, non-professional purposes, nor to a system developed solely for scientific research and not placed on the market.
3. A developer or operator established outside the Federation whose artificial intelligence systems are used within the Federation above a threshold prescribed by regulation shall appoint a representative established in the Federation, who shall be answerable to the Office and the Department of Justice for the obligations under this Act.

PART II
FEDERAL ARTIFICIAL INTELLIGENCE GOVERNANCE

Article 5: Federal Council on Artificial Intelligence.

1. There is established a Federal Council on Artificial Intelligence.
2. The Council shall be chaired by the Vice-Premier and shall include the Secretaries of Research and Development, Justice, Trade and Industry, the Interior, and Defense, and such other members as the Council of State may appoint.
3. The Council shall coordinate the artificial intelligence policy of the departments, advise the Council of State, and prepare the National Artificial Intelligence Strategy under Article 6.

Article 6: National Artificial Intelligence Strategy.

1. The Council shall prepare a National Artificial Intelligence Strategy at least once every three years and lay it before the Cortes Federales.
2. The Strategy shall set out the Federation's objectives for the development of artificial intelligence, the training of specialists, the provision of computing capacity, and the protection of the public.

Article 7: Federal Office for Artificial Intelligence Safety.

1. There is established, within the Department of Research and Development, a Federal Office for Artificial Intelligence Safety.
2. The Office shall:
   1. issue technical standards and guidance for the safe and transparent use of artificial intelligence;
   2. operate a voluntary scheme for the certification of artificial intelligence systems;
   3. administer the regulatory sandbox under Article 16;
   4. receive and examine reports of serious incidents involving artificial intelligence;
   5. conduct the Federation's safety evaluation of advanced artificial intelligence systems.
3. The Office may issue binding orders within its functions, and shall refer serious breaches of this Act to the Department of Justice.

Article 8: Dissolution of the Royal Commission on Artificial Intelligence.

1. The Royal Commission on Artificial Intelligence is dissolved, and its functions, records, and staff are transferred to the Council and the Office as the Council of State may direct.

PART III
SAFETY, TRANSPARENCY, AND HIGH-IMPACT ARTIFICIAL INTELLIGENCE

Article 9: High-impact artificial intelligence.

1. An artificial intelligence system is high-impact where it is used in any of the following fields:
   1. the supply of energy;
   2. healthcare and medical devices;
   3. financial services and the granting of credit;
   4. biometric identification;
   5. critical infrastructure;
   6. decisions by a public body that affect a person's rights or access to public services;
   7. the screening of candidates for employment or education.
2. The Office may, by regulation confirmed by the Cortes Federales, add a field to or remove a field from the list in this Article.

Article 10: Duties of operators of high-impact artificial intelligence.

1. An operator of a high-impact artificial intelligence system shall:
   1. maintain a process to identify and reduce the risks the system presents;
   2. keep documentation sufficient to explain how the system works and how it is used;
   3. provide for human oversight of the system's operation;
   4. on request, give a person affected by a decision of the system a plain explanation of that decision;
   5. carry out an impact assessment before the system is first deployed, and keep it current.

Article 11: Transparency and generative artificial intelligence.

1. Content produced or substantially altered by a generative artificial intelligence system shall be marked, by a method the Office may prescribe, so that it can be recognized as such.
2. A person shall be told when they are interacting with an artificial intelligence system rather than a person, unless this is obvious from the circumstances.

PART IV
PROHIBITED AND RESTRICTED PRACTICES

Article 12: Prohibited practices.

1. The following are prohibited:
   1. placing on the market or using an artificial intelligence system that deploys techniques beyond a person's awareness to materially distort their behavior in a manner that causes harm;
   2. the use by a public body of an artificial intelligence system to score or rank individuals by their personal characteristics or conduct for general purposes;
   3. the creation or distribution of synthetic intimate imagery depicting a real person without that person's consent.

Article 13: Remote biometric identification.

1. Real-time remote biometric identification of persons in publicly accessible spaces is prohibited.
2. The prohibition in this Article does not apply to the Federal Gendarmerie or the intelligence services of the Federation where they act under a warrant issued by a federal judge, on one of the grounds set out in regulation, and for no longer than the warrant allows.

Article 14: Synthetic media and elections.

1. Synthetic media depicting a real person or event shall be clearly labeled as synthetic.
2. The distribution of unlabeled synthetic media depicting a candidate, official, or party in connection with an election is prohibited.

Article 15: Biometric and health data.

1. An artificial intelligence inference drawn from a person's biometric or health data shall not be used to set the terms of insurance, to refuse insurance cover, or to screen a person for employment.
2. An operator of an artificial intelligence system that processes biometric or health data shall obtain the person's consent and shall collect no more data than the system requires.

PART V
DEVELOPMENT AND COMPETITIVENESS

Article 16: Regulatory sandbox.

1. The Office shall operate a regulatory sandbox in which a developer may test an artificial intelligence system under the Office's supervision, on conditions adapted for the test and for a fixed period.
2. The Office shall establish a fast-track route to market for artificial intelligence systems it assesses to be of low risk.

Article 17: National Compute and Data Initiative.

1. There is established a National Compute and Data Initiative, administered by the Office, which shall:
   1. make computing capacity available, on shared terms, to startups, universities, and public researchers;
   2. establish a framework for access to public data for the development of artificial intelligence, subject to the protection of personal data.
2. The Initiative shall be funded within a ceiling set by the Cortes Federales, released in tranches authorized one at a time.
3. A project receiving support under the Initiative shall publish a resource and productivity statement before support is approved.
4. The statement shall identify the expected computing and energy demand, the productive or public purpose of the system, and the safeguards against using artificial intelligence merely to produce synthetic reports, surveillance, or managerial scoring without demonstrable benefit.

Article 18: Talent and skills.

1. The Council of State shall establish a fast-track route to residency for persons with qualifications or experience in artificial intelligence that the Federation requires.
2. The Department of Education, Sports, and Culture, with the universities, shall provide for the training and retraining of specialists in artificial intelligence.

Article 19: Artificial Intelligence Co-Investment Fund and tax incentives.

1. There is established an Artificial Intelligence Co-Investment Fund, which shall invest alongside private capital in artificial intelligence firms established in the Federation.
2. The Fund shall be funded within a ceiling set by the Cortes Federales, released in tranches authorized one at a time.
3. A tax credit shall be available, on terms set by the Department of Treasury, for qualifying research and development in artificial intelligence.
4. A project receiving investment from the Fund shall publish a resource and productivity statement before support is approved.
5. The Office shall report annually to the Cortes Federales on compute allocation, energy demand, regional distribution, and the measurable public or economic value of projects supported under the Initiative and the Fund.

PART VI
SUPERVISION, ENFORCEMENT, AND PENALTIES

Article 20: Supervision and administrative measures.

1. The Office shall supervise compliance with this Act, and shall act through guidance and certification before resorting to penalty.
2. The Office may order a developer or operator to bring a system into compliance, and may order a system suspended or withdrawn where it presents a serious risk or breaches this Act.

Article 21: Penalties.

1. A breach of a prohibited practice under Part IV, or a breach of the duties on high-impact artificial intelligence that causes harm, shall attract an administrative fine set in proportion to its gravity, and may be referred to the Department of Justice for prosecution.
2. A breach of any other provision of this Act shall, on a first occasion, attract guidance and a period to comply before a fine is imposed.

Article 22: Appeals.

1. A decision of the Office or the Department of Justice under this Act may be appealed to the federal courts.

PART VII
CLOSING PROVISIONS

Article 23: Commencement and application.

1. This Act shall apply to all of the Federation of Nouvelle Alexandrie.
2. This Act shall not become law unless it has been given Royal Assent by H.M. King Sinchi Roca II.
3. This Act shall be published in the Federal Gazette and made publicly available within seven (7) days of Royal Assent.
4. Part IV shall come into force on Royal Assent.
5. Parts I, II, and V shall come into force on 1.I.1758 AN.
6. Parts III and VI shall come into force on 1.I.1759 AN.
7. In the event any provision or part of this Act is found to be invalid or unenforceable, only that particular provision or part so found, and not the entire Act, shall be inoperative.

Amendments

The Government amendment to Articles 17 and 19 was tabled by Alexander Melas during Federal Assembly debate on 27.VIII.1757 AN. It required projects supported through the National Compute and Data Initiative or the Artificial Intelligence Co-Investment Fund to publish resource and productivity statements before approval, identifying computing and energy demand, the productive or public purpose of the system, and safeguards against artificial intelligence being used only for synthetic reports, surveillance, or managerial scoring without demonstrable benefit. It also required annual reporting to the Cortes on compute allocation, energy demand, regional distribution, and measurable public or economic value.

The Speaker received the amendment as a sponsor amendment on 28.VIII.1757 AN. When debate closed on 29.VIII.1757 AN, no objection had been maintained, and the bill was ordered to final division on the amended text.

Voting record

Federal Assembly final reading

The Federal Assembly passed the amended bill on 1.IX.1757 AN by 687 votes to 0, with 62 abstentions or absences recorded.

Vote summary by party

Party	Aye	Nay	Abstain/absent
Federal Humanist Party	372	0	0
Democratic Socialist Party	196	0	0
Wakara People's Party	0	0	20
Federal Consensus Party	119	0	0
Civic Governance Alliance	0	0	10
Independents	0	0	32
Total	687	0	62

Chamber of Peers final reading

The Chamber of Peers concurred with the bill before Royal Assent on 1.IX.1757 AN.

Royal Assent

The Act received Royal Assent on 1.IX.1757 AN and came into force on the same date.

Analysis

The final division passed without recorded opposition. The recorded Aye vote included the Federal Humanist Party, the Democratic Socialist Party, and the Federal Consensus Party. The Wakara People's Party, the Civic Governance Alliance, and independent deputies were recorded under the abstention or absence column.

See also

• Science and technology in Nouvelle Alexandrie
• Aldurian Economic Diversification and Recovery Act, 1755
• Royal Commission on Artificial Intelligence
• Foundation 1768
• Alexandrium