Tech Law Analysis: Chile’s Plan to Regulate Artificial Intelligence

Written by León Lanis V. , Paralegal

Artificial Intelligence (AI) is something that has been heavily discussed in the news and among expert forums. Many suggest that AI has many benefits, but that there are great dangers that may negatively impact society; some of those perils are: biased information, loss of workforce, software-based racial discrimination, amongst others. The unregulated market has also led to various lawsuits, especially regarding intellectual property. Many countries have sought to regulate artificial intelligence in order to have common standards to mitigate risks and try to classify them.

In 2021, the European Parliament started debating how to effectively regulate artificial intelligence in order to boost its benefits but allay the hazards of its implementation. Very recently the European Parliament finished said process and issued the first ever regulatory guideline for AI, known as the AI Act. This law categorises the different implementations of AI, the risks associated with such implementations and general principles to follow in order to effectively assuage its negative impacts.

Chile, around the same time as the EU started discussing the AI Law, issued the first “Chilean Government Guideline for Artificial Intelligence”, which set the guiding principles for such creations, to include: AI must serve humans (human-centric principle), sustainability purposes, inclusivity-centric, ethics and globalisation focused. This policy paved the way for the drafting of a bill, which is now being discussed in Chile’s Congress. In this blog, we will highlight the main aspects of the bill and how it may impact the future of companies.

WHAT IS AI?

Under the AI bill, the concept is defined as any software that is developed using the following techniques:

1. Automated learning strategies
2. Logical and knowledge based strategies (logical-symbolic)
3. Statistical strategies, Bayesian estimation, search and optimization methods.

Any company or creation which falls under the aforementioned developments, may trigger the risk categorisations.

RISK CATEGORISATION

Following the steps of the EU, the Chilean bill gives different risk categories based on the AI model and its outcomes. Here, the bill states two intrinsic categories and we will categorise everything not contained within those two as ‘others’.

1. Unacceptable risk AIs: the future law states that the following uses of AI are considered unlawful
   1. Those using subliminal techniques in order to manipulate and alter the behaviour of a person in the manner that it creates physical or psychological hazards.
   2. Those which take advantage of the vulnerabilities of a specific group of people due to their age, physical or mental disabilities ir order to create physical or psychological harm
   3. The use of AI from public authorities or in their representation to socially classify people by their behaviour or characteristics -actual or predictable-
   4. The use of remote identification and biometrics AI in public spaces, except for the following cases as authorised by a judge and only used by the police:
      1. Search and rescue of possible crime victims, specially children
      2. The prevention of a known and specific imminent danger
      3. The detection, localisation, identification and finding of criminals or suspects of a crime
2. High risk AIs: these will be considered a tolerable high risk implementation of AI which will be subject to special consideration and evaluation by the institutions that we will explain later on this blog. The following are considered high risk:
   1. Biometric identification in private spaces
   2. Those employed in the management of water, electricity and gas
   3. Those applied for the access to schools and universities but as well to evaluate students
   4. Those related to employment and selection processes
   5. AIs applied to employee evaluation
   6. Access to public services related AI
   7. Credit scoring
   8. Those employed for emergency cases
   9. Those employed for crime prediction
   10. Customs control AI
3. Others: any other use cases are not considered within this bill, therefore, considered lawful but subject to control.

NATIONAL COMMISSION FOR ARTIFICIAL INTELLIGENCE

In order to ensure the compliance of this law, the bill states that the Ministry of Science, Technology, Knowledge and Innovation must create a government agency entrusted with the authorisation, monitoring and sanctioning of AI creations and companies.

The AI Commission will have the following roles:

1.  Ruling on the applications for authorisation of AI developers, suppliers and users,
2. Advise the government on national strategies for AI
3. Elaborate yearly reports on the state of the art
4. Create and maintain a national registry of AI
5. Rule over hazards, incidents and malfunctions of registered AIs

AUTHORISATION FOR THE DEVELOPMENT OF AI

As aforementioned, every developer, supplier or user wanting to develop, commercialise, distribute or apply AI systems will need to submit the technology to an authorisation before the AI Commision.

The commision will proceed to check the level of risk of the system. The developer, supplier or user subjected to the process will have to include every technical documentation and explanation in order for the Commission to successfully grant the authorisation. The whole process should not take longer than 60 days, subject to modifications in case of any observations.

Within this authorisation, the AI Commission will ask any creation classified as high risked to comply with the following requisites:

1. Create a risk management and mitigation plan
2. Data management plan
3. Count with an automatic event registration
4. Count with a quality management system
5. Supply users with manuals and user’s guides
6. Have enough techniques and technologies in order to effectively monitor the functioning of the AI
7. Demonstrate an adequate level of precision, solidity and cybersecurity
8. Actively monitor non authorised access of malicious third parties to the system

PENALTIES  

Not complying, in the future, with the aforementioned may result in the following penalties:

1. Not complying with the procedure of authorisation or simply not presenting the creation for authorisation will result in a 200 UTM fine (roughly USD $14.000)
2. The development, implementation or commercialisation of any unlawful AIs may result in jail time.

CONCLUSION

Although this bill was presented very recently, it has received positive reception from experts in the subject. This is yet to be discussed by both chambers (Congress and Senate) and will probably undergo many changes. Nevertheless, it is very likely to keep many things the same, especially in regards to the creation of the AI Commission and the authorisation process.

The introduction of the AI bill in Chile marks a significant step towards regulating the development and implementation of artificial intelligence technologies within the country. With its emphasis on mitigating risks and ensuring ethical usage, the bill sets forth a framework that aligns with global efforts to address the challenges posed by AI. The delineation of risk categories, establishment of the National Commission for Artificial Intelligence, and the authorization process underscore Chile’s commitment to responsible AI governance. While the bill is subject to further deliberation and potential modifications, its core principles signal a proactive approach to harnessing the potential of AI while safeguarding societal interests. As Chile progresses towards enacting comprehensive AI legislation, it sets a precedent for other nations to follow in fostering innovation within a framework of accountability and ethical stewardship.

Harris Gomez Group METS Lawyers ® opened its doors in 1997 as an Australian legal and commercial firm. In 2001, we expanded our practice to the international market with the establishment of our office in Santiago, Chile. This international expansion meant that as an English speaking law firm we could provide an essential bridge for Australian companies with interests and activities in Latin America, and to provide legal advice in Chile, Peru and the rest of Latin America. In opening this office, HGG became the first Australian law firm with an office in Latin America.

As Legal and Commercial Advisors, we partner with innovative businesses in resources, technology and sustainability by providing strategy, legal and corporate services. Our goal is to see innovative businesses establish and thrive in Latin America and Australia. We are proud members of Austmine and the Australia Latin American Business Council.