Argentina: Introduced Bill 2505-D-2023, establishing a Legal framework for the Regulation of the development and use of AI, including algorithm design and technical standards

On 8 June 2023, Bill 2505-D-2023 establishing a Legal framework for the Regulation of the development and use of artificial intelligence (AI), was introduced in the Chamber of Deputies of Argentina. The Bill would require the development of AI systems to follow fundamental ethical principles, including respect for "human dignity, privacy, transparency, responsibility and fairness", and prohibit the use of AI for illegal, discriminatory, malicious purposes or violating human rights. The use of data for development training will require users' informed consent, and the users will have the right to request explanations of the decisions made by AI systems. The Bill establishes a liability regime, the developers, providers and users of AI systems would be responsible for the consequences and decisions related to the use of AI. The developers would be required to ensure that their systems are safe, reliable and implement adequate security measures to prevent unauthorised access, tampering, or malicious interference and conduct risk assessments to identify their level of risk. The Bil establishes four risk levels, "unacceptable risks" are systems that could severely impact the individuals' fundamental rights or the security of people, and appropriate mitigation measures will be established, or the system will be banned. The "high risks" are systems considered to have a significant impact on fundamental rights or the safety of people, and "limited risks" have a minor impact on fundamental rights, and both would require mitigation measures. The systems that have a minimal impact on fundamental rights or the safety of people will be considered "insignificant risks", and no additional requirements will be established for them. Prohibition or restriction of use may be enforced when systems are found to violate human rights, engage in unfair discrimination, pose serious security risks, manipulate or unduly influence individuals or processes, lack transparency and accountability, or perpetuate inequalities or injustice.