Chile: Principles and axes of Chile's AI Policy
Artificial Intelligence ('AI') has been gradually acquiring an important role in people's daily life and businesses, which is why, in Chile, regulating and encouraging its use and development is needed. In fact, Chile has recently published its National Policy on Artificial Intelligence1 ('the AI Policy'), which addresses the benefits of AI technology and the goal of positioning the country as a leader in innovation, research, and development, and of democratising technology solutions. Macarena Gatica L., M. Ignacia Ormeño Sarralde, and Jaime Urzúa, respectively Partner, Associate, and Associate at Alessandri Abogados, analyse Chile's AI Policy and break down its key principles and axes.
In 2020, the Ministry of Science, Technology, Knowledge and Innovation called for the citizens' participation in the process of working on the three main pillars for a future AI policy on the national level. Thereon, multiple roundtables were organised with academics, as well as representatives from the private and public sector. As a result, the AI Policy was published on 27 October 2021. Thus, Chile joined other Latin American countries, which have established similar actions, such as Argentina, Brazil, Mexico, and Colombia.
The scope of the AI Policy is to provide a ten-year guideline with the objective of empowering people in the use and development of AI technology in Chile, and of integrating technological advances into the realm of well-being and human development. This way, the AI Policy aims to insert Chile in a global collaboration environment related to AI, within an ecosystem of research, development, and innovation, intended to contribute to a sustainable development and improvement of the quality of Chilean life.
The AI Policy is structured around four cross-cutting principles and three interdependent axes. Each axis addresses the opportunities and gaps in its field, whilst introducing objectives and priority actions that the country must undertake in a time horizon of ten years (i.e. in 2031).
In addition, the AI Policy includes an AI Action Plan, which brings together 70 priority actions with public investment that will impact areas such as education, productive development, and talent training. The Action Plan specifies the initiatives included in each of the priority actions of the AI Policy, addressing responsible parties and execution deadlines within the next ten years.
Indeed, in the second semester of 2021, the Ministry of Economy, Development and Tourism has chosen AI for the first regulatory sandbox. This sandbox aims to create regulatory standards, allowing the development of innovation and anticipating the undesirable effects, such as rights violations or bias.
In addition, the Chilean Economic Development Agency ('CORFO') called for proposals aiming to diminish productivity gaps identified in prioritised productive sectors through the development, adjustment, and implementation of AI systems. This program consisted of CORFO providing an amount of around $1,200,000 in the form of a non-refundable subsidy covering up to 60% of the total cost of the program.
Finally, Chile is undergoing a drafting process of its new Constitution, expected to end in 2022. In this context, there is a significant interest regarding digital transformation, with fundamental rights in the digital environment being a priority, such as the universal right of access to internet, digital literacy, personal data protection, cybersecurity, critical infrastructure, and, above all, an ethical and adequate use of AI.
Transversal principles of the AI Policy
AI with a focus on people's welfare, respect for human rights, and security
AI must contribute to the overall well-being of the people. Actions must be oriented towards providing people with a better quality of life, taking advantage of AI benefits, and addressing AI risks and potential negative impacts, whilst respecting the human rights of all citizens.
Algorithms and the data used to train automated systems, especially when dealing with personal data, must be reliable. Therefore, these must include risk and vulnerability assessments to avoid a use of data that could lead to arbitrary decisions or results, and to protect critical algorithms that, if breached, could compromise Chilean security.
AI for sustainable development
This principle focuses on AI as a potential avenue for emerging countries to diversify their economic matrix, making industries more productive, encouraging research, technology, innovation, and commercial applications.
The actions must promote the use and development of technology, strengthening the ecosystem, and incorporating AI as an axis of sustainable development for Chile, with social and environmental considerations.
Data integrity and quality should be used to recognise and address possible bias. In addition, this principle emphasises the need to develop AI in an inclusive, non-discriminatory manner, promoting gender, cultural, and sexual diversity, and inclusiveness of historically neglected groups, such as people with disabilities.
AI should generate regional and/or macro-zonal strategies based on the distinctiveness of each territory and according to specific benefits that AI may produce for a particular area.
Globalised and evolutive AI
AI should position Chile in line with the treaties subscribed by Chile, promoting its participation in bilateral and multilateral agreements: for example, the Organisation for Economic Co-operation and Development's ('OECD') Principles on AI and the OECD AI Council, or the Digital Economy Partnership Agreement ('DEPA') between Singapore, New Zealand, and Chile.
Axes of the AI Policy
This axis focuses on talent development, technological infrastructure, and data.
- Talent development refers to providing adequate knowledge, skills, education, and training in data, mathematics, statistics, engineering, computer science, and computer programming, among others. These skills and knowledge should be taught from primary through to higher education as part of the educational curriculum and course syllabi. Moreover, AI must be a cross-discipline in vocational, technical-professional training for a broad spectrum of workers. This is expected to increase the number of AI professionals.
- In relation to technological infrastructure, an adequate connectivity, accessible platforms, and data centres are, inter alia, necessary to improve Chilean productivity and development. Regarding internet access, Chile stands out against the rest of Latin America and the Caribbean, with more than 82% of the population with access to internet, making it one of the countries with the highest coverage with a Digital Adoption Index of 75.62 according to the World Bank Group.2 Nevertheless, this percentage includes internet access through mobile telephones. The technological infrastructure should be aligned with the goal of Chile becoming a global technological hub for the Southern hemisphere.
- Data is an indispensable element for AI. Therefore, it is important to promote trust and best practices, which, in turn, enables organisations and/or companies to share data for the common good and in compliance with data privacy. The object of this axis is to promote and implement a public data agenda, including legal certainties and clear definitions of responsibilities within Chile, as well as a public-private data ecosystem, considering access to quality data for AI by different actors.
Development and adaptation
This axis includes the development of basic and applied research, technology transfer, innovation, entrepreneurship, and improvement of public services. Its objects are:
- generating AI productivity indicators;
- strengthening Chilean research and development in AI to achieve a level equal to or higher than the OECD average;
- promoting collaboration between academia and the productive sector for research and development of AI systems;
- promoting and stimulating productivity around AI to reach levels equal to or higher than the average economic growth of OECD countries;
- modernising the public sector through AI; and
- adapting to and mitigating key challenges, like climate change, with AI.
Ethics, regulatory aspects, and socio-economic effects
This axis includes ethics, impacts on labour force, consumer relations, creations, intellectual and industrial property, cybersecurity and cyber defence, and gender.
Regarding ethics, the use of AI presents risks associated with fundamental rights violations, such as dignity, privacy, freedom of speech, and non-discrimination. Therefore, it is key that AI development considers the assessment and mitigation of these risks, by implementing AI in a responsible manner.
In terms of impact on work, the technology development is changing the nature of work, requiring new generations with new skills to enable AI and to adapt to the digital era. The risks of certain tasks commonly performed by human beings will decrease, when performed by machines, thus leading to the loss of work for many individuals. Because of this, the work force must adjust to the growing interaction between humans and machines.
The massification of e-commerce is accompanied by the increasing use of AI by suppliers in consumer relationships. The goal is to promote a transparent, non-discriminatory, and respectful use of personal data.
Cyber attacks are continuously increasing and growing significantly in complexity. AI is a key element to stay at the forefront in the use of cybersecurity technologies and procedures. Therefore, AI must be a relevant component in the field of cybersecurity and cyber defence, promoting secure technological systems. In the field of research, innovation, and technology, there are significant gender gaps, horizontally and vertically. Thus, it is necessary to implement policies driven by gender equality, in line to promote the participation of women in AI until reaching at least a level equal to or higher than the OECD. Finally, promoting gender equality in the implementation of AI systems is required.
Areas of concern and discussion
Despite the recent publication of the AI Policy, there is still some way to go and areas to improve and expand upon.
On the one hand, regarding talent development, Chile suffers at all levels of skill and talent gaps related to digital transformation. There is a need to reconvert parts of the workforce. Today, professionals with knowledge in AI are scarce, and have specialised mainly through work experience or by attending courses.
On the other hand, it is necessary to update the regulation data privacy, aiming to reduce the current void in the matter. Legal certainty can foster a legal framework and achieve international harmonisation, specifically in relation to the standards set by the General Data Protection Regulation (Regulation (EU) 2016/679) ('GDPR'). Chilean-American Chamber of Commerce ('AmCham') and Universidad del Desarrollo ('UDD') conducted a survey in 2019, and concluded that 78% of the surveyed companies did not use the data they had at their disposal to make decisions and were far from implementing AI systems based on such data.3
Similarly, a survey on the access and use of information and communication technology in companies by the Chilean Ministry of Economy, Development and Tourism found that only 2.2% of the surveyed companies used techniques or tools for Big Data analysis.4
Moreover, it is necessary to increase the public-private cooperation so that both sectors can join forces to contribute to AI knowledge and deployment. Regarding access to data, it has been discussed the possibility of a data-pool, where the government could set an example, making data available to the private sector in accordance with the law.
Finally, regarding e-commerce, the AI Policy addresses the risks linked to the asymmetry existing in consumer relationships. AI may be able to generate unfair or arbitrary results, hinder transparency in the conditions for contracting of products and services, or process consumers' personal data inadequately. Consequently, suppliers that may use AI must ensure that these risks are mitigated, while encouraging the use of AI in e-commerce in a transparent, non-discriminatory manner and respecting data protection regulations.
1. Available at: https://minciencia.gob.cl/uploads/filer_public/bc/38/bc389daf-4514-4306-867c-760ae7686e2c/documento_politica_ia_digital_.pdf (only available in Spanish)
2. See at: https://publications.iadb.org/publications/english/document/Artificial-Intelligence-for-Social-Good-in-Latin-America-and-the-Caribbean-The-Regional-Landscape-and-12-Country-Snapshots.pdf
3. See at: https://amchamchile.cl/wp-content/uploads/2019/09/L.-Bravo-UDD.pdf (only available in Spanish)
4. See at: https://www.economia.gob.cl/2020/07/08/encuesta-tic-2018.htm (only available in Spanish)