During the 2023-2024 legislative periods, various bills have been introduced in both chambers of the National Congress of Argentina, the Senate and the House of Deputies. These bills are currently under discussion in committees and aim to address the diverse challenges posed by artificial intelligence (AI). For any draft to be debated by the chambers in full, a majority of committee members must first vote in favor of the text of such draft.
The proposed legislation identifies responsible management, transparency, security, and the protection of users and their personal data as guiding principles.
Some bills aim to translate these principles into specific measures, such as incorporating watermarks or disclaimers in AI-generated content (a proposal by Senators Juez and Huala) or amending the Criminal Code to penalize the dissemination of AI-generated content depicting child sexual abuse (proposals by Deputies Lospennato and Brügge).
Other lawmakers propose more comprehensive approaches, inspired by the European Union’s AI Act. For instance, legislators such as Brügge, Ávila, and Romero suggest classifying AI systems based on their risk levels and banning their application to certain practices.
Among the prohibited practices outlined in the bills proposed by Brügge and Romero are: a) manipulation of behavior that may cause harm or expose individuals to risk, b) social scoring systems resulting in discriminatory treatment or denial of rights, and c) real-time biometric identification in public spaces, except under specific circumstances like tracking fugitives, locating victims, or preventing terrorist attacks. Notably, Brügge’s proposal is more restrictive than Romero’s and even the EU’s AI Act, as it requires prior judicial authorization for these exceptions.
While several lawmakers draw inspiration from the European regulatory framework, significant concerns arise regarding the risk-based approach. Unlike the AI Act, which classifies AI systems and imposes proportionate obligations based on risk levels, many of the proposed legislation in Argentina fail to consider such classifications. This omission results in indiscriminate regulatory requirements for developers and users of AI systems. For example, the bills by Brügge, Ávila, and Romero mandate registration with oversight authorities and/or conducting impact assessments for all AI systems, regardless of their specific risk levels. Such requirements are alarming, as they may impose unnecessary regulatory costs, hindering innovation and progress, particularly in developing economies. Imposing indiscriminate regulatory burdens on AI-centered enterprises could result in prohibitive financial costs, adversely affecting projects where these technologies pose no significant risk to citizens’ rights and freedoms. Over-regulation could also harm the general population by limiting access to beneficial technological advancements.
Another critical issue is equipping oversight authorities—proposed in several bills—with the technical, human, and financial resources necessary to fulfill their supervisory and auditing responsibilities for AI systems.
Deputy Yeza’s proposal stands out for adopting a different approach from the EU AI Act and the afore mentioned bills. Yeza suggests creating a regulatory sandbox system, which would serve as a testing ground for enterprises applying AI and other emerging technologies to solutions in key sectors such as health, education, security, and agriculture. The proposed law includes monitoring these developments to assess their performance, impact on stakeholders, scalability, and the need to adapt Argentina’s existing regulatory framework to enable their implementation.
The sheer number of bills introduced in Congress undoubtedly reflects a genuine interest among legislators to promote the ethical and responsible use and development of AI while mitigating risks to individuals and society. These risks include misinformation, bias, violations of fundamental rights such as privacy, environmental harm, and threats to democracy, among others. Although some proposals adopt questionable approaches and pose significant implementation challenges, they can be optimized through constructive and serious debate in Congress. Such discussions must involve all relevant stakeholders, including technology companies, civil society, subject-matter experts, privacy and human rights professionals, academics, and representatives of affected communities.
We hope the coming years will witness inclusive and profound debates where all voices are considered, ultimately leading to a regulatory framework tailored to the specific needs of Latin America, and particularly Argentina. This framework should balance technological development, innovation, competition, protection of individual rights and freedoms, ethical principles, and democracy. In this regard, initiatives like Yeza’s regulatory sandbox offer a valuable opportunity to foster multi-stakeholder dialogue and enable oversight authorities to better understand the phenomenon they are tasked with regulating.
