Page 25 - AI Standards for Global Impact: From Governance to Action

AI Standards for Global Impact: From Governance to Action AI Part 1: International

Figure 16: Sasha Rubel, Head of Public Policy for Generative AI, Amazon Web Services
(AWS) High-level roundtable - From principles to practice: How AI standards enable
effective governance

Q4� What role can standards play in ensuring AI supports inclusive, sustainable
development?

• Participants advocated for global AI standards as equity enablers to promote algorithmic
fairness, explainability, and non-discrimination; emphasized that standards can serve as
tools for inclusion through participation in standards development, particularly for the
Global South, while supporting interoperability and contextual relevance; highlighted
standards' value as drivers of responsible innovation, helping to ensure sustainable
scaling by managing risks like privacy and security; and positioned standards as a shared
language for cross-border, cross-sector coordination, stressing the need for inclusive, co-
created frameworks aligned with sustainable development goals and ethical foresight.
• To help ensure AI inclusivity, said participants, open-source models could be paired
with bias and sustainability metrics to account for differences in country-specific training
data, supported by regulatory sandboxes and active collaboration across academia,
governments, and the broader ecosystem.
• Participants highlighted the potential future importance of large-scale AI models,
suggesting that conflicts might target AI systems rather than traditional infrastructure,
underscoring the need for standards that balance global coordination with national
sovereignty.
• Participants highlighted the need for standards to support data governance and help
ensure data quality (e.g. using sensors for measurement), improve AI model training
efficiency, optimize resource distribution, and enable the reuse of existing datasets. They
also emphasized the need for creating standards to measure AI energy consumption and
defining interfaces to help small companies, researchers, and industry partners integrate
with large AI platforms effectively.

13