Chile submits AI regulation bill to congress

The Chilean government presented a bill to regulate artificial intelligence that follows the logic of the regulation approved by the EU with different obligations depending on the risks.

A draft regulation in Peru also establishes obligations according to risks.

The Chilean initiative, which entered the lower house on May 7, defines unacceptable risk systems as those that are incompatible with the respect and guarantee of people's fundamental rights, which is why their use would be prohibited. Meanwhile, it also defines high-risk systems as those that present a significant risk of causing harm to health, safety, fundamental rights or the environment, as well as consumer rights.

Limited risk systems are defined as those that present an insignificant risk of manipulation, deception or error, as a result of their interaction with people. AI systems without obvious risk are also defined.

UNACCEPTABLE RISK SYSTEMS

Among the unacceptable risk systems, the bill mentions manipulative systems, except those for authorized therapeutic purposes and with informed consent, and systems that exploit people's vulnerabilities to generate harmful behaviors.

Regarding biometric identification systems, the regulation prohibits the use of AI for such purposes in public spaces in real time, although the prohibition will not be applicable in the case of public security and criminal investigations.

The bill would also prohibit AI systems based on facial scraping techniques through the non-selective extraction of facial images from the internet or images from closed circuit television and systems intended for the evaluation of a person's emotional state.

The proposed legislation also establishes an obligation to comply with data governance and cybersecurity.

In the case of limited risk systems, the proposed regulations provide for transparency mechanisms such as clear and precise information on the interaction with artificial intelligence, with the exception of criminal investigations.

An AI advisory council, which would be created by the law, will have the responsibility of presenting a list of high-risk and limited-risk AI systems.

A personal data agency would oversee and ensure compliance with the law.

The full bill can be downloaded in the Documents box in the top-right corner of the screen.