Peru advances with EU-inspired artificial intelligence regulation

Peruvian lawmakers have started debating a proposal to create a framework for the use and development of artificial intelligence (AI) technologies, which is largely inspired by the European Union’s AI rules.

Submitted this week by lawmaker Carlos Javier Zeballos Madariaga, bill 07033/2023 is the first in the country to specifically tackle AI technology. Other countries across the region, such as Paraguay and Brazil, are also advancing with AI regulations.

The Peruvian framework is intended to “guarantee the protection of essential rights, the privacy and security of citizens, and promote transparency, ethics, innovation and progress in the field of artificial intelligence,” the summary of the proposal states.

The bill’s author also says that another goal is to “avoid risks and negative consequences” of the technology.

The bill advances the concept of data sovereignty by stipulating that the rules will apply to any company or person that develops or operates AI in Peruvian territory, regardless of where the servers supporting the technology are located.

In other words, even if AI technology is hosted on datacenters in other countries, it will still be subject to regulation if it is offered and applied in Peru.

Data sovereignty is the idea that data are subject to the laws and governance structures of the nation where they are collected.

AUDITING, TRANSPARENCY AND MORE

AI developers will have to document and disclose how algorithms operate for external auditing purposes, something that is also a key pillar of EU legislation.

In line with Peru's personal data regulation, the AI bill also states that collection and processing of users' data can only take place with clear and explicit consent.

Furthermore, individuals will also be able to request explanations about how the AI works and what their data will be used for at any time.

Meanwhile, mobile apps that allow users to browse or download services based on AI technology that does not comply with the rules and principles of the legislation will be subject to "drastic" sanctions.

Like the EU rules, Peru's legislation sets out different risks related to the use of AI systems, with specific rules and restrictions for each case yet to be defined. 

In the bill, four types of AI risks are being considered: unacceptable, high, limited and insignificant risks, depending on their impact on essential or human rights. 

In the case of unacceptable risks, the recommendation is for a full ban of the technology.

The bill also makes the government and digital transformation secretariat, which is under the control of presidency of the council of ministers, responsible for overseeing the legislation and enforcing its rules.

The full text of the bill can be seen here, in Spanish.

The text will have to be debated by the congressional science, technology and innovation committee, as well as by the budget and general accounting committee, before it can be discussed and voted on by the full congress.