The UN Secretary-General’s AI Advisory Body, convened by Secretary-General António Guterres last October, launched its Interim Report: Governing AI for Humanity at the beginning of January calling for feedback. The report outlines AI opportunities, risks, and enablers, and proposes to strengthen international AI governance by carrying out seven critical functions built on five Guiding Principles. These principles emphasize AI inclusivity and public interest, the centrality of data governance, a collaborative approach across stakeholders, and alignment with the UN Charter, Human Rights Law, and UN SDGs.
Below are a few conceptual remarks for the definition of Institutional Functions:
AI Regulations & Incentives. AI “guardrails” together with multi-level incentive strategies are essential to move to a human-centric and sustainable development and scale-up of AI. They should work in synergy and be driven by shared values and global societal needs, rather than by fast-evolving technologies. To realize this transformation, it’s crucial to highlight the business and innovation immense potential of developing AI solutions with high social and environmental impact and increase its attractiveness through multi-level incentive strategies.
AI Sustainability. The document does not address environmental, social, and economic sustainability aspects of AI uptake. Additionally, long-term risks associated to behavioral and business changes of AI uptake and growing climate effects have not been considered (e.g., rebound effects). Moreover, the need to move to a sustainable and ethical design of AI systems, which integrates environmental, social, and economic sustainability considerations into algorithms and models and evaluates short and long-term system impacts, has not been discussed. Past experiences have shown the inefficiency and drawbacks of the widespread approach “develop first and add sustainability considerations later”.
AI opportunities. In the report little attention has been paid to the AI potential for fostering individual growth and human capabilities beyond cognitive skills, which should be included into the guiding principle. The development of AI systems should also be driven by impact, thus resulting in a more positive connotation though this explicit reference in the guiding principles “AI systems should play an active role in fostering human growth, both at the individual and community level, while reinforcing human uniqueness and diversity through their functionalities. Human qualities should never be portrayed as subordinated or in competition with machines; it should be clearly communicated that AI is a human artifact for societal advancement.”
Tech-centric vs. sustainable human-centric approach. It’s crucial to clarify in the report that the success of AI doesn’t depend just on the ability to develop new technologies but on the ability to do that with a human-centered and sustainable mindset, considerate of short and long-term impacts. Changing mindset and considering the bigger picture with short and long-term implications for all stakeholders and the environment is the biggest challenge we face, which requires setting priorities for the development of AI and multi-level incentives.
You can find the review of the Interim Report below.