Peru's National Superintendency of Customs and Tax Administration (SUNAT) has approved its policy for the safe, responsible and ethical use of systems based on artificial intelligence (AI). The framework establishes twelve commitments for applying AI in tax and customs management, including human oversight, data protection, traceability, bias prevention, transparency, risk management and respect for fundamental rights.
The decision sets out the rules that must guide the development, implementation and use of AI systems within Peru's tax and customs administration, an especially sensitive area given the nature of the information SUNAT manages and its direct relationship with taxpayers, companies and customs operations.
Peru's AI Law reaches SUNAT
The policy is framed within Law No. 31814, which promotes the use of artificial intelligence in support of the country's economic and social development, and its regulations, approved by Supreme Decree No. 115-2025-PCM. The legislation requires Peruvian public entities to have an institutional policy on the matter.
In the case of SUNAT, the proposal was drafted by a designated technical team and subsequently approved by the institution's Government and Digital Transformation Committee during its session on June 16, 2026. This committee will also be responsible for implementing, monitoring and evaluating compliance with the new policy.
The approval of the document therefore means transferring the general framework established by Peru for the development and use of artificial intelligence to the operation of a specific institution that is especially relevant to the country's economic activity.
In its policy, SUNAT recognizes systems based on artificial intelligence as institutional assets and enabling technologies for digital transformation and the continuous improvement of tax and customs management. The institution thus presents artificial intelligence not merely as a technological tool, but as part of its digital transformation process.
The purpose set out in the document is to align the use of AI with SUNAT's Institutional Strategic Plan, while guaranteeing the protection of fundamental rights, the confidentiality of information, tax secrecy and the trust of taxpayers. The combination of technological innovation and information protection occupies a central position in the policy, especially given the volume and sensitivity of the data handled by a tax and customs administration.
SUNAT's 12 commitments for using artificial intelligence
The annex to the resolution sets out twelve institutional commitments that must govern the development, implementation and use of systems based on artificial intelligence within SUNAT. The first is compliance with the current legal framework. The institution establishes that the use of these systems must conform to the Constitution, Law No. 31814, the Tax Code and the General Customs Law, while promoting well-being, equity and inclusion.
The second commitment consists of directing artificial intelligence systems toward generating public value, with the aim of improving the services provided to taxpayers and increasing the efficiency of tax and customs management. The third establishes an ethical, responsible and people-centered approach based on respect for human rights and dignity.
Human oversight of AI systems
The fourth commitment establishes that there must be human oversight of systems based on artificial intelligence, so that the results produced by these technologies can be reviewed, corrected or halted where appropriate. The policy thus introduces human intervention as one of the principles that must accompany the institutional use of AI.
The fifth commitment relates to traceability and accountability. SUNAT must establish traceability, audit and continuous improvement mechanisms that make it possible to hold accountable the actors involved in the development and use of these systems. The institution must also pay attention to the possible discriminatory effects of algorithms.
As a sixth commitment, SUNAT undertakes to prevent, mitigate and correct biases, errors or discriminatory results, seeking to ensure that the use of artificial intelligence is proportionate and consistent with the purposes for which it has been implemented.
Personal data and confidentiality
The handling of information is another of the major blocks of the policy approved by Peru's tax administration. The seventh commitment thus expressly establishes the obligation to protect personal data, tax secrecy and the confidentiality of information, in accordance with applicable regulations.
This is one of the especially relevant aspects of the new framework due to the nature of the information managed by SUNAT. The eighth commitment refers to technological security. The artificial intelligence systems used by the institution must be secure, reliable and robust. To achieve this goal, SUNAT establishes the application of a risk management approach throughout the entire lifecycle of AI systems.
Risk management must therefore accompany these technologies from their development through to their use within the institution's different processes. The ninth commitment establishes that the entity must promote transparency and the explainability of the operating criteria of systems based on artificial intelligence, albeit always within the limits established by legislation.
The policy thus incorporates the need to be able to explain the operation of the systems used by the institution as one of the elements of its artificial intelligence governance model. The tenth commitment refers to intellectual property. SUNAT expressly establishes that it must respect copyright and related rights during the development and use of systems based on artificial intelligence.
The new policy also addresses the impact of artificial intelligence on the institution's workers and the need to prepare them for the use of these technologies. The eleventh commitment establishes that SUNAT must train its staff in ethics, risk management and data protection, fostering an institutional culture of responsible innovation.
Training thus becomes one of the tools envisaged for translating the general principles established by the policy into the everyday use of artificial intelligence. The last of the twelve commitments extends the analysis of the effects of these technologies beyond their technical operation.
SUNAT must assess the human, social, economic, cultural and environmental impacts of artificial intelligence, while also seeking its alignment with the Sustainable Development Goals (SDGs).
AI in auditing, taxpayer services and customs control
The approval of this policy positions SUNAT among the first Peruvian public entities to formalize its own framework for the use of artificial intelligence in compliance with Law No. 31814. The relevance of the document is directly related to the institution's remit.
The next step will be to translate the twelve commitments set out in the policy into the day-to-day operation of the tax and customs administration. Implementation will be in the hands of SUNAT's Government and Digital Transformation Committee, which will be responsible for monitoring and evaluating compliance with the policy.
The committee must translate the twelve commitments into concrete procedures, controls and training within the institution. It will be in this implementation phase that principles such as human oversight, traceability, bias prevention, protection of tax secrecy, transparency and risk management must become mechanisms applicable to the operation of artificial intelligence.