COLUNAS

  1. Home >
  2. Colunas >
  3. Humanidades e Novas Tecnologias >
  4. AI Compliance and Environmental Responsibility: Bridging the Gap

AI Compliance and Environmental Responsibility: Bridging the Gap

quinta-feira, 16 de maio de 2024

Atualizado às 13:24

There is considerable discourse surrounding the impacts of AI, with current research indicating its potential to challenge all fundamental rights. However, the issue of climate and environmental justice, and what legislation and compliance can do to mitigate risks in this area, are significantly lagging behind. There is a predominant focus on extolling the benefits of the 6th wave of technological innovation and the new industrial revolution, often treating foundational models as if we were to repeat what happened in previous waves of innovation according to Schumpeter's concept of "creative destruction," all the while disregarding the exponential speed at which we live and the global impacts of AI. While some research already suggests a paradigm shift, moving from a "human-centered AI" ethos to one that is "life-centered," in order to consider not only the impacts on individual rights but also collective and social ones, and to rethink the concept of "smart cities" and "innovation," questioning for whom and for what, associating ethics (meta-innovation) and responsibility for innovation (Luciano Floridi and Wolfgang Hoffmann-Riem), practical efforts are still timid, and concrete measures to turn abstract principles into concrete practices are lacking.

On the other hand, considering the principle of prevention, once environmental damage occurs, it is almost irrecoverable, meaning it's difficult to return to the "status quo ante". In the area of data protection, such notion is equated by Omri Bem Shahar, addressing data pollution, as well as highlighting the preventive function, for example, by Eligio Resta and in the proposals of the EU's AI-ACT and the bills in Brazil, regarding compliance measures, such as the algorithmic impact assessment, but still with various flaws, starting with its non-mandatory nature, in addition to the lack of standardization and establishment of fundamental requirements. Hence, what may happen in practice is what is called ethical washing or makeup practices, without objective and concrete changes, such as the case of the Myanmar's genocide, with Facebook being held responsible by human rights judicial system, even though it developed a Human Rights Impact Assessment, but after the fact, that is to say, without fulfilling its essential function, which is to prevent harm before it occurs.

An approach must be adopted, then, for the study and practical alternatives to the issues raised that are inclusive, interdisciplinary, and holistic, moving away from the "mantra" "be fast and break things." As it is known, there is a relationship between the concepts of epistemic, environmental, data, algorithmic, and social justice, however, the impacts and benefits of AI are not distributed equitably between the Global South and North, and that's why Roberto Mangabeira Unger points out the need in countries of the Global South to decolonize thought and think universally from the local perspective (i.e., observing the social-cultural context), which involves concepts of multiculturalism as emancipatory and progressive, intercultural as proposed by Boaventura de Sousa Santos, focusing on diversity and recognition of cultural difference underlying the idea of interculturality, multiplying participation and cooperation spaces - politics as cultural politics.

Thinking critically about the role of law, which is necessarily linked to the realization of Justice, highlights risk prevention management and new forms of sociability, and reinforces the role of law in maintaining the Rule of Law instead of understanding it as a "commoditization of Law" (Richard Susskind, Daniel Susskind), avoiding what is called systemic corruption or corruption of law - legal illegality (K. Marx, Montesquieu, Marcelo Neves, Niklas Luhmann), avoiding that we only have what Derrida points out as "force of law," remaining after the erasure of the word law, only force, when absolute politicization would occur, and total degradation of law to a sort of disguise of politics, a mere instrument of power, forgetting that the idea of law, the "spirit of laws," however, is justice.

Thinking about this relationship between the different forms of justice, we point to some developments in the area of concepts and scope of rights as it happened with privacy, supported by the concept of collectivization, recognizing its substantial and procedural dimension and the right to an informational procedure, abandoning the individualistic, anthropocentric, thomistic, and proprietary perspective (Stefano Rodotá, Luciano Floridi, Helen Nissenbaum, Alessandro Mantellero, Danilo Doneda), such as the "group privacy theory" of Luciano Floridi/University of Oxford - after the Cambridge Analytica case and the theory of privacy related to "contextual integrity" of Helen Nissenbaum, in addition to the "right to reasonable inferences" as a new right of data protection, as postulated by Sandra Watcher and Brent Wittelstadt, and the "right to procedural regularity" (Joshua Kroll).

Hence, talking about sustainability by design and the Rule of Law since its conception and fundamental rights by design, addressing the central issue of environmental impact, emphasizing the UN's Sustainable Development Goals (SDGs) related to the environment, social justice, and sustainability, and in connection with the G20's agenda and priorities, focusing especially on sustainability, inclusion, and social justice, specifically with the values and objectives of the sheriff trails, "environmental and climate sustainability". Therefore, there is a relationship that cannot be disregarded between climate sustainability, inclusion, social justice, and environmental justice, a justice-oriented approach to ensure that the AI ecosystem and its benefits are distributed more equitably. The talk about data colonialism arises, and in the environmental area, the concept of "carbon colonialism" emerges, meaning that rich countries in the Global North are outsourcing the impacts of their resource extraction to the poorer countries of the Global South, in addition to the theme of the invisibility of workers from the south ("ghost workers"), like those who remove toxic plastics from landfills - a key role in ending plastic waste. It is pointed out that waste flows to the Global South, while capital flows to the Global North (Laurie Parsons, Carbon Colonialism: How Rich Countries Export Climate Breakdown). Although the topic is urgent and of utmost importance, there is still a lack of legislative proposals focused on the life-centered AI perspective and environmental impacts, for example, the two main bills aimed at regulating AI in Brazil (PL 2120/2338/23), the first without any mention of the environment and the second, although it mentions, does not establish anything more concrete. Similarly, the GDPR and the EU's AI ACT, in their initial version, are silent, and although the latest version of the AIACT from June 2023 brings some improvements, they still depend on voluntarism and abstraction.

Therefore, we must consider the correlation between AI, equity, data colonialism, and climate justice, especially because the damages of climate change are distributed unevenly, requiring equitable participation in strategies to combat climate change. Furthermore, we have to think about how compliance and the tool of algorithmic impact assessment could help prevent such damages, thinking about the development of a more protective "framework" than the currently existing ones, such as Nist and ISO standards, and since such theme and elaboration will involve the analysis of fundamental rights norms in collision, there will be a need for professionals in the area of fundamental rights theory, such as the legitimate interest assessment document in the GDPR of the European Union and in the LGPD of Brazil brings a necessary consideration to be made; it is also urgent to rethink the entire epistemological structure of elaborating such theory by Robert Alexy, as it is not adapted to the context we currently live in the "onlife" environment (Luciano Floridi), based on his discussions with Mart Susi ("Proportionality and Internet", 2020) and evolving since they did not address the impact of Artificial Intelligence or the environmental issue.