An Open Letter to the Global South:
Bring the “rest” in
During the last few centuries, human society has undergone several transformations. Different technological innovations have allowed real revolutions to permanently change the way we live, relate to each other, and understand the world. Better tools, steam engines, penicillin, vaccines, telecommunications, plastic, silicon, computers.
The so-called 4th Industrial Revolution is the culmination of the digital age, where technologies such as robotics, nanotechnology, genetics, and artificial intelligence (AI) promise to transform our world and the way we live. And currently, the most accessible and massively used of these innovations is AI.
You don't have to look very far to find autonomous systems interacting with you or the people around you. Your cell phone probably operates several different types of models trained by machine learning (e.g., language models, computer vision), among other methodologies that fall under the big interdisciplinary umbrella we call Artificial Intelligence.
Drones deliver our packages, recommendation algorithms shape and influence our preferences, probabilistic models optimize various processes previously reserved only for the human intellect. Autonomous cars, protein folding, natural language processing, climate models, every year (or month) a new barrier is overcome.
As much as “intelligence” is an elusive concept to define, there are more than a few who believe that we are on a path toward genuine artificial intelligence. True AI.
Intelligence promises to produce some of the most urgent solutions humankind needs. Clean energy sources, cures for disease, better economic systems. All these solutions are “golden eggs,” and as Ray Solomonoff, one of the founding fathers of the artificial intelligence research field, said: the AI research field is not looking for specific golden eggs. We are after the goose that produces them.
Genuine AI may well be the last invention we need to make. And even if the path to artificial general intelligence is blocked by a still unknown barrier, this does not mean that weak AI will change, as it is changing, the world in a direction that is still, as the future always is, uncertain.
And this is nothing new.
Many organizations and global actors, foreseeing the urgency and importance of this issue, are already taking action to ensure that this uncertain future is a worthwhile one. Given the transformative power of AI, whether for the benefit or detriment of humankind, it is of great importance that we reach a consensus on the standards and guidelines that will regulate such technologies.
From this sense of urgency, several research fields have emerged, such as AI Ethics and AI Safety. Both with their specific interests and agendas, but with a common overall goal: To make the interaction between humans and AI safe and beneficial. To ensure that Artificial Intelligence will be created for the good of all.
Technology is neither good nor bad. However, it is not neutral. Technology, like AI, is what we make of it. Thus, we cannot use our creations to excuse ourselves for our mistakes. In a world where models are created, trained, developed, deployed, and used by people, the answer “the data is biased” will not be enough to protect and preserve our values.
To ensure that our values are in line with the goals being optimized by AI systems, and to ensure that the values that guide the industry responsible for developing such systems are not just the values of the market but values that promote the flourishing of our species, various ethical guidelines have been published in recent years.
This is a very important initiative, and we encourage and congratulate everyone involved in this endeavor. In these published documents, with their most diverse origins, many ethical principles are defined and defended: Transparency, Justice, Equity, Non-maleficence, Accountability, Privacy, Beneficence, Freedom, Autonomy, Trust, Dignity, Sustainability, Solidarity.
The purpose of this letter is in no way to criticize the work done so far. Ethical guidelines are written by people. People are historical entities with specific contexts. And it is only natural that the principles outlined in such guidelines reflect the preferences of those responsible for producing them. The purpose of this letter is a request that two ethical principles, seemingly overlooked by many such guidelines, receive more attention in this debate: Diversity and Inclusion.
Diversity: The valuing of the different ways in which the human entity can come to express itself by whatever group or identity it wishes. AI systems must be developed in a way that protects and values our diversity.
Inclusion: The embracing and welcoming of all the ways that the human entity can come to express itself, regardless of specific affiliations, groups, and identities. AI systems should be developed in such a way as to "include," not exclude.
The literature in AI ethics has already advanced this debate to the point of creating several definitions of “algorithmic justice,” which are nonetheless ways of implementing principles such as diversity and inclusion. However, the same cannot be said about the field of AI Ethics and AI Governance itself.
This is not so much a claim as an empirical fact about the current state of the world. The vast majority of the ethical guidelines and documents created to regulate and govern AI and its industry have their origins tied to organizations in the Global North.
Virtually the entire global South is outside the debate about the ethical principles that should regulate and direct the future technological development of our society. And this is unacceptable.
To our knowledge, the only ethical reinvindications made by Latin America are the 2019 “Declaración de Principios Éticos Para La AI de Latinoamérica”, the 2020 “fAIr LAC: Adopción ética y responsable de la inteligencia artificial en América Latina y el Caribe”, and the 2020 “Marco Ético para la Inteligencia Artificial en Colombia”. In the Middle East, the only ethical claim regarding AI is “Dubai's AI Principles” of 2019. In South Asia, we only find the “Compilation of existing AI ethical principles (2020)” and the “Proposed Model AI Governance Framework: Guiding Principles (2019)”, both from Singapore. In Sub-Saharan Africa, we only found the “AI & Data Topical Guide Series” from 2020. Of the nearly 200 ethical guidelines published to date, these are the only voices from the Global South. And by no means do they represent the totality of social complexity that lies below the equator.
Regions like Latin America, Sub-Saharan Africa, the Middle East, and South Asia still have a very small voice in this debate. One cannot expect other countries and cultures to define what is best for the context of these places. Places marked by deep scars from the old world, by colonial wounds and imperialist policies that will not be healed anytime soon.
This letter is a plea for the Global South to wake up. And a request that the Global North be ready to hear it and to receive it. If we really want AI to be “a warm light for all mankind to share,” we cannot forget that we live in a plural, unequal, and diverse world. We must remember the voices that until now, have not had the opportunity to claim their preferences, explain their contexts, and who knows, tell us something we have not yet considered or heard.
To all global actors in the North, do not forget that the North/South divide is illusory. We all share the same house. And to all the new and potential actors in the global South, make your voices heard. Use whatever communication vehicles are within your reach. Engage and claim your rights.
Bring the “rest” in.
Nicholas Kluge Corrêa, AIRES (AI Robotics Ethics Society) PUCRS
Nythamar de Oliveira, PUCRS - CNPq
Diogo Fernando Massmann, PUCRS
Carolina Del Pino Carvalho, PUCRS
Lara Sosa Márquez, PUCRS
Camila Palhares Barbosa, PUCRS
James William Santos, PUCRS
Camila Trindade Galvão, PUCRS
Rodrigo Mambrini Sandoval Barbosa, PUCRS
Edson Antônio Sousa Pontes Pinto, FCR - PUCRS
Edmund Terem Ugar, University of Johannesburg (UJ)
Rodrigo Leal de Moraes, NAVI Tecnopuc
Guilherme Camargo, PUCRS
Aaron Hui, AI Robotics Ethics Society
Guilherme Fernandes Garcia, NAVI Tecnopuc
Cláudio Teixeira Damilano, PUCRS
Alexandre de Souza Athaíde, PUCRS
Renata Lima Ferreira, Centro Universitário Ritter Dos Reis
Délio José Kipper, ESMED-PUCRS
Márcia Cassimiro, Fiocruz/IOC
Rafaela Weber Mallmann, PUCRS
Tiago Protti Spinato, Unijui
Matheus de Mesquita Silveira, UCS
Denise Cantarelli Machado, PUCRS
Ricardo S. dos Santos, PUCRS
Hellen Maruyama Kodama, Universidade do Vale dos Sinos
Antonia Wallig, Associação Cultural Vila Flores
Diogo Sebastião Mury, Lenda Cripto
Camile Costa, AI Robotics Ethics Society
Renata Guadagnin, Terceiro Andar Educacional; PUCRS
Nelson Fossatti, PUCRS
Érico João Hammes, PUCRS
Nuno Miguel Pereira Castanheira, PUCRS