LAMPROS Stergioulas
UNESCO Chair on Artificial Intelligence and Data Science for Society
Urgent global action on AI governance towards commonly and widely agreed frameworks and standards is a key priority, and there are useful paradigms of technology-focused international cooperation which can be useful in this endeavour, such as ITU. Further complacency and delay are not an option, and it may well be needed to proceed rapidly with shaping global AI governance ecosystems and to initiate cooperation actions, even without waiting for a complete set of global standards first to be developed via the traditional institutional routes.
Responsible AI has not yet been fully achieved. The central obstacle is that AI governance remains largely at the level of principles and conversations and has not yet been translated into systematic implementation and scaled adoption. Trust is the driver of adoption and the condition for AI to generate benefits at scale for society. Healthcare provides a clear example: successful pilots do not automatically translate into adoption across entire health systems. Robust governance and safeguards must be built into both design and implementation before innovation can be scaled. AI governance currently faces two major bottlenecks. The first is the breakneck speed of technological development, which governance institutions are struggling to match. Governance must move beyond dialogue towards collaboration and, more importantly, substantive international cooperation. The second is the growth of inequality and injustice within the global AI system, which may further marginalize certain countries and communities as the technology spreads. Human beings must remain in control. The future should not be determined simply by what available tools can do, but by what kind of society humanity wants to design. Global governance should therefore accelerate the construction of international ecosystems for sharing, open exchange and the development of common standards.
