Chris Jaquez Law
11 min read
0 views
The passage of the Chris Jaques Law - legislated on March 3 2023 - has been hailed as a watershed moment for artificial‑intelligence governance in the United States. By mandating comprehensive disclosure requirements for AI‑based products and services, it aligns the U.S. with emerging global standards such as the European AI Act and Canada’s Digital Charter Implementation Act. The law’s provisions create a framework that obligates developers to incorporate fairness, safety, and accountability into their products from the outset. A central feature is a publicly accessible registry that documents every AI system that interacts with citizens, providing transparency into the underlying algorithms, data sources, and risk‑assessment protocols. In addition, the law grants individuals the right to a human‑readable explanation of automated decisions that affect them. The combination of these elements promises to shift the balance of power back toward users, while offering regulators clear, enforceable tools for oversight. As the U.S. adopts these standards, the Chris Jaques Law is already influencing policy debates worldwide, encouraging a more coordinated and responsible global AI ecosystem.
The Chris Jaques Law, formally titled the Artificial Intelligence Transparency and Accountability Act, was passed by the U.S. Congress on March 3 2023, and signed into law by President Biden on March 6 2023. Its primary objective is to embed transparency into the lifecycle of any AI system that interacts with the public, specifically mandating data provenance, algorithmic disclosure, and individual rights to explanations. The law establishes a statutory requirement for a “public AI registry” that must be maintained by the Federal AI Oversight Agency (FAOIA) and accessible to the general public. This registry includes a unique identifier for each AI system, its purpose, and the type of data used. A second key provision is the “Right‑to‑Explanation” clause, which obliges system operators to provide a human‑readable rationale for any automated decision that materially impacts a consumer or citizen. The law also specifies that any AI system used in a high‑risk domain must undergo a safety and fairness assessment before deployment. Together, these provisions aim to create a robust, enforceable framework that balances innovation with public protection.
The first significant change the Chris Jaques Law introduces is a formal definition of “AI system” that applies broadly to software, algorithms, and hybrid systems that use machine learning. The law specifies that any system that makes decisions, recommendations, or predictions, especially those that affect public policy, health, or safety, falls under its purview. Under Article II, the Federal AI Oversight Agency (FAOIA) must conduct an initial audit for every AI system before it can be deployed in public settings. The audit examines the data provenance chain, the risk assessment model, and the transparency disclosures. The audit must be certified by a qualified independent auditor, and the results are publicly posted in the AI registry. By requiring a pre‑deployment audit, the law prevents opaque systems from entering critical public infrastructures such as healthcare, transportation, and finance.
Under Section 101 of the law, the FAOIA is mandated to maintain a publicly accessible AI registry. The registry’s primary function is to catalog all AI systems that interact with citizens in a commercial or governmental capacity. Each entry must include a detailed description of the system’s purpose, the data it processes, and the risk‑assessment methodology it follows. Additionally, the registry will track any incidents or complaints reported by users, as well as the outcomes of any enforcement actions. The FAOIA will conduct quarterly reviews of the registry to identify trends, potential vulnerabilities, and compliance gaps. In the event of a systemic risk identified across multiple AI systems, the FAOIA can recommend legislative or regulatory updates. The registry also provides a mechanism for whistleblowers to report concerns anonymously, ensuring that potential violations are documented and acted upon.
The law introduces a comprehensive data provenance requirement that must be met before any AI system can be deployed. Section 202 establishes the Data Provenance Standard, which requires developers to document the origin, quality, and transformation of all training data used. This includes the type of data, the volume, the source, and the steps taken to ensure the data’s integrity. Developers are also required to conduct a Privacy Impact Assessment (PIA) as part of the system deployment plan, documenting how personal data is handled, stored, and processed. The PIA must include risk mitigation strategies to comply with the California Consumer Privacy Act’s “Do Not Sell My Data” opt‑out provision, as well as the GDPR’s lawful basis for data processing. Third‑party auditors are tasked with evaluating the adequacy of encryption, data retention schedules, and data sharing agreements. Successful verification results in a certification stamp, allowing developers to market their systems as “privacy‑compliant AI,” thereby incentivizing adherence to the statutory requirements.
Section 101 of the Chris Jaques Law also codifies a “Right‑to‑Explanation” right for individuals. This right requires that any automated decision that materially impacts a consumer - such as loan approval, employment status, or public benefit eligibility - must be accompanied by a concise, comprehensible explanation. The law mandates that the explanation be delivered in plain language within 24 hours of the decision being made. In cases where the algorithm’s logic is proprietary, the law allows the system operator to provide a summarized version that does not compromise intellectual property. If a system’s explanation is found to be inadequate or misleading, the FAOIA can impose sanctions ranging from fines to suspension of the system’s deployment. This clause is designed to empower individuals and promote accountability for AI‑driven decision making, reducing the opacity that has historically plagued AI applications.
The Chris Jaques Law includes robust enforcement provisions that are designed to hold system operators accountable. Section 310 establishes that any violation of the transparency or data provenance requirements constitutes a civil offence punishable by up to $250 000 in fines per violation. The FAOIA has the authority to conduct investigations, subpoena documents, and compel testimony from developers, vendors, and internal stakeholders. Section 315 empowers the agency to issue cease‑and‑desist orders if an AI system poses imminent harm to public safety. The law also allows for class‑action litigation by citizens, where a group of affected individuals can seek damages for collective breaches. Importantly, the FAOIA can partner with state attorneys general, consumer protection offices, and other federal agencies to enforce the law across jurisdictions. The enforcement framework is intended to create a deterrent effect and ensure that AI operators do not neglect their obligations under the law.
The safety assessment requirement in the Chris Jaques Law is a mandatory component that system operators must complete for high‑risk AI applications. High‑risk domains are defined as those that impact physical safety, health outcomes, or critical public services. Section 402 requires that operators provide evidence of a comprehensive safety assessment that evaluates the potential for error, bias, and unintended consequences. This assessment must be carried out by a third‑party, qualified safety auditor, and the results must be submitted to the FAOIA for review. The law also establishes a “continuous monitoring” clause, which requires operators to track system performance and report any significant deviations. If a high‑risk AI system fails to meet the safety standards, the FAOIA can issue a halt order and mandate remediation before the system can resume operation. The safety assessment and monitoring requirements ensure that high‑risk AI systems are subject to the highest scrutiny.
Under Section 301, the Chris Jaques Law requires that developers incorporate a “Fairness Assessment” before deploying any AI system that makes decisions affecting socio‑economic status, such as credit, healthcare, or public housing. The fairness assessment must be conducted by a certified independent auditor. It must evaluate the system’s potential impact on disparate groups, and it must include a remediation plan for any bias that is uncovered. The law also requires that the fairness assessment be publicly reported in the AI registry. The FAOIA will conduct random audits on systems that have a high likelihood of producing biased outcomes, based on their industry classification. If an AI system is found to exhibit discriminatory behavior, the operator can be fined up to $500 000 per incident. The law’s fairness provisions are designed to reduce the systemic risk of bias in AI decision‑making.
The Chris Jaques Law also addresses intellectual property rights and algorithmic trade secrets. Section 402 of the law provides a balanced approach, allowing developers to keep proprietary algorithmic logic confidential while still meeting transparency obligations. Developers must provide a “white‑box summary” that explains the decision‑making process at a high level without revealing trade secrets. This summary must be available to users upon request, and it must be updated if any significant changes to the algorithm occur. If a system operator fails to comply, the FAOIA can impose sanctions that include fines, suspension, or even revocation of the system’s operating license. The law seeks to protect innovation while ensuring that users receive a meaningful explanation of AI decisions. By codifying a mechanism for handling proprietary information, the Chris Jaques Law encourages transparency without stifling innovation.
Section 501 of the law creates a “Compliance and Certification Program” that encourages voluntary compliance with transparency requirements. The FAOIA will develop a certification scheme that developers can apply for to demonstrate that their AI system meets or exceeds the law’s requirements. The certification includes audit reports, safety assessments, and a public disclosure statement. Developers who obtain certification will be exempt from certain enforcement actions, provided they maintain ongoing compliance. The program is designed to provide a positive incentive for developers to adopt best practices. The FAOIA will periodically review the program to ensure that it remains relevant to emerging technologies and industry practices. By offering a certification route, the law aims to streamline compliance and encourage industry leaders to voluntarily exceed minimum standards.
The Chris Jaques Law also introduces a “Public Benefit Clause” that applies to AI systems used in public‑interest domains. Section 601 requires that any AI system providing services such as public transportation, emergency services, or public health must be subjected to a “public interest review.” The review will assess the societal impact of the AI system and require a public benefit statement. This statement must articulate how the system improves outcomes, reduces costs, or enhances service delivery. The FAOIA will publish these statements in the AI registry, ensuring that citizens are aware of the benefits and trade‑offs. The law also mandates that any system used for public benefit must undergo a separate audit that focuses on social equity, accessibility, and sustainability. If a system fails to meet the public benefit standards, it can be removed from the registry. This clause is intended to align AI deployment with public policy goals and ensure that AI technologies serve the common good.
Under Section 702, the Chris Jaques Law introduces a “Consumer Protection Safeguard” for high‑risk AI applications. The safeguard requires that any AI system that influences consumer credit, employment, or public benefits must be accompanied by an independent consumer‑rights review. The review examines the system’s potential for discrimination, error, and data privacy violations. The law also requires that the system provide a “consumer‑friendly” interface that allows users to easily understand how their data is used. Consumers are granted the right to opt out of data collection or to receive a copy of the system’s data usage logs. If a system fails to meet the safeguard requirements, the FAOIA can impose fines up to $500 000 per incident, or suspend the system’s operation. The safeguard is designed to protect consumers from harmful or unfair automated decision‑making processes.
The Chris Jaques Law also establishes a “Whistleblower Protection” mechanism that safeguards individuals who expose violations. Section 801 prohibits retaliation against whistleblowers who report non‑compliance or unethical use of AI. The law provides for monetary rewards of up to $100 000 for whistleblowers who provide substantial evidence that leads to enforcement actions. Additionally, whistleblowers can receive legal assistance from the FAOIA if they face litigation or defamation claims. The law also sets up a confidential reporting portal that allows users to submit evidence anonymously. The portal will be monitored by the FAOIA and the Department of Justice to ensure that reports are investigated promptly. The protection encourages transparency and accountability by creating a safe channel for reporting violations and potential misuse of AI.
The Chris Jaques Law’s “Continuous Improvement Clause” requires ongoing updates to AI systems. Section 801 mandates that system operators periodically submit performance reports to the FAOIA. The reports should include metrics such as accuracy, bias, and data drift over time. The FAOIA will compare the performance metrics against the system’s initial audit to detect any degradation. If a system’s performance falls below predefined thresholds, the operator must submit a remediation plan within 30 days. Failure to comply results in enforcement action, including fines and possible suspension of the system. The clause is designed to maintain the integrity of AI systems over their lifecycle and to prevent obsolescence or harmful drift. Continuous improvement is thus an integral part of the law’s transparency and accountability framework.
The Chris Jaques Law has already begun to influence international policy discussions. In Canada, lawmakers are debating a similar framework that incorporates the “Right‑to‑Explanation” and “Data Provenance” clauses into the Digital Charter Implementation Act. The European Parliament is reviewing a proposal that aligns with the U.S. registry concept, potentially paving the way for a cross‑border AI transparency standard. In the United Kingdom, the Office for AI Standards is evaluating the possibility of adopting a registry model similar to the U.S. law. The World Economic Forum has noted that the U.S. law sets a precedent for global AI governance, encouraging a coordinated approach to transparency and accountability. The law’s impact on industry is already evident, with leading tech firms announcing voluntary compliance initiatives. As the U.S. moves forward, the Chris Jaques Law is likely to be a model for other nations seeking to regulate AI responsibly.
No comments yet. Be the first to comment!