By Paola Cantarini
This text is part of my participation as a moderator at the event held on 17 and 18.10.24 in Fortaleza, VI SIAJUS, “Effects of informational asymmetry on society and private relations”, an initiative of UNI7, AID-IA and the Ethikai Institute, and is part of my post-doctoral research at USP/RP, with a FAPESP grant, where I present part of the conclusions of my research.
Despite a growing discourse on the ethical implications of AI, as can be seen in the work of Heleen Janssen, Michelle Seng Ah Lee, Jatinder Singh, and the initiatives of important bodies such as the European Commission and UNESCO, there remains a critical need for frameworks that holistically integrate fundamental rights and environmental considerations. That’s why I’d like to thank Professor Kevin for his contribution, as he points out that the AI hype can distract us from the problems that are already present, and that this is also a constant concern pointed out by various researchers, such as the Alan Turing Institute, where I also work in one of the research groups.
The hype surrounding AI could increase the risks in the cases mentioned of the use of chatgpt, used, for example, by a lawyer to draw up a legal petition without reviewing or checking the data, creating non-existent case law. Hence the importance of a critical vision as a possible third way, as opposed to a utopian or dystopian vision of AI, both of which are excessive and blind to the other side. This is why we emphasize the need for a critical, interdisciplinary and holistic approach, since we consider the origin of AI with cybernetics to be a holistic discipline that is more appropriate for analyzing complex scenarios. In post-modern times, the problems are unprecedented and there are no ready-made formulas or answers. It’s a work in progress. And we need to broaden perspectives and collaborations beyond academia, democratizing the discussion and making it more inclusive, including broadening the dialogue between North and South, since most scientific production and initiatives are still centered in the global North.
In addition, although recent documents, such as “Climate Change & AI: Recommendations for Government Issued” by the Global Partnership on AI, have begun to recognize the environmental dimensions of AI, practical methods for mitigating these impacts remain largely unexplored. It is essential to seek to fill this gap by innovating in areas where previous research is limited, offering an innovative approach to ensure that the development of AI is consistent with human dignity, social equity and ecological balance, and fundamental rights.
This is why we call for a broader look, going beyond privacy by design and default to fundamental rights by design, in order to consider the potential impact of AI on all fundamental rights, considering the environmental impact. (Human rights in national AI strategies Source: Bradley et al., 2020; “Getting the future right”, European Union Agency).
If we look at the initiatives in Germany’s strategic vision for AI, as well as in Brazil and the US proposal and the EU’s AI ACT and from Brazil, although they stress moving in a direction that benefits the common good, putting people first and discussing a human centric AI approach, this approach is still insufficient and needs to be extended to a life centered AI perspective. Although the “German” AI ethics perspective on AI stresses that the ethical and legal issues of AI go hand in hand, concrete steps need to be taken to turn ethical agendas and principles into effective measures, avoiding practices known as “ethical laundering” (Luciano Floridi), and likewise “compliance laundering”. It is therefore important to anticipate potential risks by analyzing them now, rather than focusing on a dystopian future of existential risks, which could distract us from the current problem. Although the perspective of “human-centered AI” is important, meaning human control of technology and respect for human values, this perspective is insufficient, as it does not take into account the environmental impact for the most part and is not aligned with a long-term sustainable approach, as it does not address the notion of the multidimensionality of DF, in the individual, collective and social aspects. As far as the approach of the European Union, Brazil and Germany is concerned, although they raise concerns about the environment, there is no mention of a sustainable approach. With regard to the approach of the European Union, Brazil and Germany, despite raising concerns about the environment, there is no reference to how this would be done, nor to carrying out an impact analysis of AI.