Normative alert: AAIP Guide for Public and Private Entities on Transparency and Personal Data Protection for Responsible Artificial Intelligence
Data Privacy and Personal Data – Cybersecurity Department Report | Normative alert: AAIP Guide for Public and Private Entities on Transparency and Personal Data Protection for Responsible Artificial Intelligence.
Dear All,
On September 27, 2024, the Government of Argentina published the AAIP’s Guide (the “Guide”) for the responsible use of Artificial Intelligence (“AI”) on its website (available only in Spanish here). The Guide aims to ensure that public and private entities using these tools guarantee the right to transparency and the protection of citizen’s personal data.
Its objective is to “address risks and concerns during the use of AI, both in processes aimed at personalizing user experiences, making evidence-based decisions, as well as those focused on automating processes, reducing time, and improving results, among others”.
The key recommendations are as follows:
-Conducting impact assessments: These allow the identification and mitigation of potential risks related to data protection before the implementation of AI-based systems.
-Assembling multidisciplinary teams: Diversifying the sectors and areas of expertise within teams facilitates addressing the technical and ethical challenges of AI.
-Compliance with the principle of explainability: AI systems must be transparent and explainable, though it is acknowledged that this may present technical challenges in some cases.
-Protection of personal data: AI systems must include security measures that ensure the correct handling of personal data and protect the rights of data subjects.
-Consideration of the complete AI lifecycle: Each stage of the AI system lifecycle requires the consideration of specific recommendations. The life cycle consists of the following four stages, defined based on OECD definitions: • System Design: In this stage, the system is planned and designed, considering the selection and processing of input data, algorithmic model design, and model training. • Verification and Validation: Once the system is defined, it undergoes testing to ensure correct functionality. • Implementation: This involves putting the system into operation. For this, it is necessary to have the infrastructure where it will be deployed. • Operation and Monitoring: In this stage, the system is fully operational, available for use, and constantly monitored for performance.
We remain at your disposal for any additional information you may need.
Yours sincerely,