The UN Secretary General’s AI Advisory Body recently published an interim report on international governance of AI. António Guterres established the group in October 2023 with a mandate to build consensus and strengthen global cooperation around AI governance. Members of the High-Level Advisory Body include Ran Balicer and Amandeep Singh Gill, two former Commissioners of the Lancet-Financial Times Commission on Governing Health Futures 2030.
Welcome focus on AI and health
The DTH-Lab has submitted its feedback on the Advisory Body’s report as part of a public consultation. We welcomed the fact that the report highlights the potential of AI to transform public health. From the very first introductory paragraph, health is identified as one of the sectors likely to be significantly impacted by AI and examples are found throughout the report of where AI models are already being used to support health goals.
Moreover, the report warns of potential health benefits and risks of AI that extend beyond its application in health systems. The major societal and environmental shifts likely to arise from AI will have consequences for all areas of life, intersecting with determinants of health to shape our health and well-being in complex ways. The future of AI raises fundamental questions about what it means to be human and, in the most extreme scenario highlighted by the Advisory Body, even poses an existential threat to humanity.
“The role of AI in disintermediating aspects of life that are core to human development could fundamentally change how individuals and communities function. As AI capabilities further advance, there is the potential for profound, structural adjustments to the way we live, work, and interact.“
— UN Secretary General’s AI Advisory Body
Strong alignment with the Governing Health Futures 2030 Commission
The Lancet-Financial Times Commission on Governing Health Futures 2030’s recommendations for stronger governance of digital transformations have been well reflected in the Advisory Body’s report. The DTH-Lab therefore acknowledged many positive elements of the report and its recommendations in our submission:
Forward looking and anticipatory: The Advisory Body is concerned not only with governance of AI today but governance that can respond to changes in technology and its uses. Governance mechanisms need to be agile and informed by ongoing horizon-scanning to identify potential consequences of AI. This includes sector-specific governance. Policymakers focused on health and other sectors must stay abreast of technological innovations and anticipate their potential impacts.
Precautionary governance: As the Advisory body states, “scientific uncertainty about risks should not lead to governance paralysis”. The DTH-Lab shares this view. In line with public health practice—and the severity of potential risks related to AI—urgent governance action is warranted whilst additional research is conducted to better understand AI’s impact on different groups.
Principled approach: Rather than proposing a single model for AI governance, the Advisory Body has developed a set of principles for international AI governance—built on norms and values enshrined in the UN charter and human rights instruments—that different governments and partners can align with. Resisting a one-size fits all approach is prudent given that many different models of AI and data governance are emerging around the world and countries are at different stages of digital maturity. Principled-based governance also aligns with the mission-oriented and value-based approach proposed by the Lancet-FT Commission. Specifically, the Commission recommends that governance of digital transformations should be founded on the Health for All values set by WHO, namely democracy, equity, solidarity, inclusion, and human rights.
Inclusive governance: The Advisory Body’s first guiding principle is that AI should be governed inclusively, by and for the benefit of all. Another principle emphasizes the importance of multi stakeholder collaboration. This is another strong area of alignment with the DTH-Lab and Commission report which argue for inclusive forms of governance that enable individuals and groups to actively participate in and cocreate the design and governance of digital transformations. In particular, children, youth, women, and marginalized communities must be at the forefront of inclusive AI governance.
Sharing risks and benefits equitably: The introduction to the Advisory Body’s report raises concerns about the benefits and harms of AI not being shared evenly as well as the risk of AI driving further inequalities. Such concerns were also highlighted in the Commission’s report, resulting in calls for investments in enablers such as digital infrastructure, connectivity and digital literacy.
Three recommendations for the AI Advisory Body
The DTH-Lab agrees with the Advisory Body’s assessment of the risks and challenges associated with AI. We particularly share the concerns raised about the potential role of AI in exacerbating inequalities, undermining democracy and further shifting power towards commercial interests.
In our submission, we propose three areas where we believe the AI Advisory Body’s guidance could be strengthened.
- Focus on maximizing the public value of AI: We encourage the Advisory Body to reframe their second guiding principle from governing AI “in the public interest” to maximizing the public value of AI. Assessing public value requires weighing up risks of an AI use case for people and communities against the public benefits that it promises to create. Whilst public interest is a binary notion, public value assessments are more nuanced and take into account the sustainability of AI’s benefits, future societal needs and the distribution of potential benefits and harms. Public value assessments also consider how to safeguard against risks and correct any harms that may occur. Our white paper on Data Solidarity provides further background on the merits of foregrounding public value in all aspects of digital and data governance.
- Highlight health as a priority sector for implementation of the Advisory Body’s recommendations. As our recent working paper outlines, AI governance recommendations to date have generally focused on its general purpose nature and not distinguished between the various uses or sectors to which AI can be applied. Due to the scale of potential opportunities and risks arising from the application of AI both within health systems and society as a whole, we believe that health should be among the first sectors in which the Advisory Body’s governance recommendations should be taken forward. As the Advisory Body looks to further develop and operationalize its recommendations, it should already begin to explore how the principles and functions of a global AI governance regime could be implemented in support of health goals.
- Take a generational and intersectional approach to AI governance. The risks and challenges and opportunities associated with AI are not uniform to all. Young people’s experiences of AI, for example, are—and will be—very different to other generations. Individual experiences of AI and other aspects of digital transformation are shaped by a wide range of social, economic and other factors. We therefore encourage the Advisory Body to consider generational, intersectional and gender-sensitive perspectives in their analysis of the opportunities, enablers, risks and challenges of AI and to recommend that any international AI governance mechanisms do the same.
With the hype around AI showing no sign of decreasing, the Advisory Body’s work to build consensus on the international governance of AI could not be more critical. We eagerly await the group’s second report in August 2024 and hope to see our recommendations incorporated. We encourage our partners to also share their inputs through the public consultation which closes on 31 March.
