Spotlight: Peru's AI regulations

A draft regulation for Peru's law 31,814 that promotes artificial intelligence is in the public consultation period through June 1 and aimed at supplementing a law from last year to establish key guidelines.

Related bills are in congress.   

The draft is available in the Documents box in the top right corner.

Law 31,814 defines the principles guiding AI as adherence to risk-based security standards, a plurality and stakeholder approach, internet governance, ethical and responsible development, and privacy.

The proposed regulation establishes that the government and digital transformation cabinet will act as enforcement authority. Other governance actors are the personal data protection authority, public administration entities, civil society organizations, citizens, academia and the private sector.

BNamericas reviews the main points of the proposal.

RISK MANAGEMENT

The regulation classifies AI systems based on risk.

Systems used to modify human behavior through deception or by taking advantage of vulnerabilities of a group that could lead to prejudice are deemed unacceptable, as are systems that classify individuals or groups by quantifying social behavior.

Regarding high-risk systems, obligations like data traceability, room for human control and security measures are established.

PRIVACY AND TRANSPARENCY

Entities that implement AI systems would be required to clearly tell users that the service or product they interact with is based on artificial intelligence. They must also communicate the parameters used to reach a decision or result, and provide information on the rights of those potentially affected.

Companies must also employ diverse and multidisciplinary teams to design, develop and implement AI-based systems, follow international best practices, ensure compliance with copyright and intellectual property law, promote awareness and avoid automation bias.

PUBLIC SECTOR

Public entities implementing AI must also employ multidisciplinary teams, foster information exchange between entities to generate evidence and databases for decision-making, and reinforce statistics, machine learning and programming skills, among others.

They will also have to coordinate with entities that manage state information and contribute to the systematization of accurate and up-to-date information via the national open data platform.  

Any algorithms, models and data processed via AI systems that received some form of state support must be made public on the open data platform and the public software platform for general use.

INNOVATION

Authorities will use the national digital innovation and AI center, created in 2020, to define actions to attract and retain talent, design and develop prioritized use cases in coordination with the competent entities and promote alliances, including international ones, with other institutions. 

-The center will also help coordinate hackathons and similar events with the government and digital transformation laboratories to promote the ethical and responsible use of AI.

In line with the regulation, a national high performance computing center will be established to provide digital infrastructure and services to public entities.