CAF – the development bank of Latin America and the Caribbean – has reaffirmed its commitment to the responsible and ethical use of artificial intelligence (AI) in all its operations and projects. This initiative aligns with that of forming a working group for the creation of a Regional AI Ethics Council for the region, announced on October 23, 2023 in Santiago, Chile, as part of an agreement signed between CAF and UNESCO.
With this vision, CAF takes on the challenge of promoting and leading the ethical and transparent adoption of AI in the institution and in countries in the region by evaluating the impact of the technology it develops and uses. It is essential that technology is a driver to build a more inclusive, equitable and just society, providing better opportunities for all.
In this declaration, a governance system and a management model have been defined which guarantee respect for ethical and responsible AI. With a scope that covers all areas and geographic areas of the Institution, and which extends throughout the life cycle of the services and projects executed, from their design and structuring to their implementation.
“At CAF, we promote inclusive, multi-sector AI governance from our internal processes, strengthen the region’s voice in global discussions and advance the development of talent, data and key infrastructure. CAF makes the necessary financial instruments available to the region to advance this program. For example, we recently approved $1.5 million for the High Performance Computing for Artificial Intelligence in Latin America and the Caribbean project, which is a crucial step toward creating a regional cognitive computing network,” said Sergio Diaz. Granados, executive president of CAF.
Within this declaration, CAF undertakes to respect the following principles:
Transparency: CAF is committed to working to make all decisions and processes based on AI transparent and understandable.
Accountability: CAF is committed to regularly monitoring and evaluating the performance of AI systems to ensure ethical and responsible decision-making focused on human participation.
Ethics: CAF is committed to using AI in a manner consistent with internationally recognized ethical standards and the principles set out in the CAF Code of Ethics.
Confidentiality: CAF will ensure that information is processed ethically and securely, respecting individual rights, complying with applicable privacy laws and regulations and protecting the information of CAF, its partners, allies and collaborators.
Fairness: CAF will ensure that the AI systems it uses are not discriminatory or perpetuate unfair biases.
Sustainable use: CAF will encourage responsible practices and investments in AI technologies that have a positive impact on mitigating the effects of climate change, as well as other key aspects of sustainability. This will encompass energy efficiency, responsible management of natural resources, social and economic equity, diversity and inclusion.
People-centered: CAF will encourage the use of AI that respects human rights and is committed to achieving the United Nations Sustainable Development Goals. Guided by these elements, we will strategically apply AI to drive innovation, address critical challenges, and move toward a more sustainable future.
Efficiency: CAF will use AI to make its operations and those of its partners and allies more efficient.
To comply with the principles mentioned above, CAF has established a governance system consisting of executive leaders and technology and ethics experts, responsible for overseeing the implementation of AI in the bank. Where CAF will establish the procedure if necessary to approve or refuse projects and tools using AI. The Digital Transformation Committee will regularly review AI-related policies and practices, ensuring they are aligned with established ethical principles and corporate strategy.
CAF will implement biannual monitoring and evaluation of AI systems to ensure they comply with established ethical principles. The Digital Transformation Committee will evaluate annual reports on the results achieved through monitoring and evaluations. The CAF will publish annual reports detailing the performance of its AI systems and the steps taken to address potential ethical challenges.
CAF will collaborate and comply with international AI-related organizations and standards.
