On June 2, 2023, resolution 2/2023 of the Undersecretariat of Information Technology was published in the Official Gazette of the Argentine Republic. This resolution establishes ethical principles and recommendations for the design, development, implementation and/or use of projects based on artificial intelligence ("AI") in the framework of the civil service. This resolution provides two annexes:

• Annex I: Introduces "Guidelines for Reliable Artificial Intelligence". This handbook gathers international precedents on the subject and equips users with tools for public innovation projects involving technology and AI. This handbook aims to offer theoretical and practical tools to those who work in the civil service, whether leading innovation projects, developing technologies, or adopting technologies developed by other technical teams/suppliers or formulating technical specifications for these acquisitions.
• Annex II: This is a graphic outline of the stages of creation, design and development of AI projects.

Annex I: Handbook and recommendations

The handbook sets out the conceptual framework applicable to the AI projects to be carried out, lists a series of recommendations for the creation, team building and control of AI projects. The AI development cycle is divided into 4 stages, namely:

1. The design and data modeling stage.
2. Verification/validation stage
3. Implementation stage
4. Operation and maintenance stage.

Conclusion

Based on the regulations analyzed, we can summarize that:

• It is oriented towards those who are part of the national civil service.
• It aims to raise awareness that the development and implementation of AI may involve both technological advances and challenges in terms of values, ethics and human rights.
• These challenges demand that the adoption of AI follow a set of ethical principles which drawn from international precedents in the field, in such a way as to maintain the protection of fundamental rights.
• It creates awareness at the moment prior to the conception of AI, by explaining the concepts of responsibility and execution. This means that when employing AI algorithms, the tasks may be delegated, but the responsibility for their execution remains with the person in charge. Consequently, the algorithm itself lacks self-determination, placing the burden of accountability on the individual overseeing its operations.
• It details the ethical aspects to be considered, along with the theoretical and practical tools to be used at each stage of the AI development cycle mentioned above.