Aspen Digital

The Global Digital Compact called for agile and adaptable cooperation on AI priorities to deliver on the promise of accelerating the achievement of the Sustainable Development Goals. In order to deliver on that promise, we must more clearly understand and articulate what direction we want to be headed in. The scientific panel and global dialogue should work together to address this question.

I appreciate the Global Digital Compact’s language around opportunity assessment. In order to achieve the SDGs, the global AI community must not only avoid the potential risks associated with AI, but also clearly articulate the priorities, challenges, and opportunities that AI must address in order to be a true public good. 

Without meaningful measures of progress, the AI industry may continue to pursue goals that may be orthogonal or even antagonistic to the SDGs. While there are many claims being made, we lack answers to basic questions. Are the AI capabilities being developed today actually going to help with gender equality (goal 5)? With food security (goal 2)? With decent work (goal 8)? And more?

This should be the primary work of the scientific panel and global dialogue.

The scientific panel should prioritize development of metrics for tracking AI impacts on the SDGs, including developing technical benchmarks that clarify whether the capabilities of AI systems are fit-for-purpose to address the important challenges that the world faces.

To achieve this, I recommend that the panel work with UN departments like the Department of Economic and Social Affairs and Secretary-General’s Innovation Lab as well as academic and civil society institutions to translate SDG indicators into meaningful metrics for evaluating and guiding the development of AI systems. Developing these metrics will require articulating the economic and technological enablers of the SDGs. It will also require meaningful assessment that tracks the impacts of AI adoption on the SDGs in different geographies, markets, and domains. The panel should meet at least quarterly and publish updates on benchmarks and other measures regularly so that meaningful public accountability can be achieved.

Similarly, the Global Dialogue has an opportunity to articulate targets for AI development. I am proud to be a collaborator in the (similarly named) Global Dialogues initiative already underway from the Collective Intelligence Project. This effort leverages deliberative technologies mentioned by previous speakers to elevate the voices of people from around the world to clarify the fears, dreams, hopes, and attitudes people have about AI.

The mandate of the UN Global Dialogue called for by the Global Digital Compact should be to identify priorities and goals for AI development. The UN Global Dialogue should incorporate input from the Collective Intelligence Project and others to ensure that these priorities and goals reflect a diverse community of global voices. This can complement and advance the work of the scientific panel to ensure that AI development actually helps accelerate the most important goals we have as a global community.

The UN has an important role to play in articulating what we want the future of AI to look like. Like the SDGs, articulating benchmarks and indicators for AI will help to steer this technology in service of global good. Now we need to identify priorities so that we can ensure that those tasked with building these tools actually focus on the right things to get us there.