Page 103 - AI Standards for Global Impact: From Governance to Action

AI Standards for Global Impact: From Governance to Action Part 2: Thematic AI

Figure 45: ITU report on measuring AI's environmental impact

Key takeaways:

• AI's environmental impact is real, complex and growing, not only in terms of energy and
emissions, but also with regard to water usage, material intensity and equity considerations.
• Significant progress has been made in energy efficiency, but further efforts are needed
to standardize measurements and expand impact assessments to include the full lifecycle
and broader ecological factors.
• There is a clear need for harmonized global standards and multistakeholder cooperation,
especially involving regulators, private sector, and UN agencies, to guide AI deployment
in alignment with climate goals.
• Opportunities exist for ITU-T Study Group 5 to advance work on AI-specific standards,
particularly in:

o Model-level environmental accounting
o Data centre efficiency benchmarking
o Lifecycle reporting frameworks
o Water use and biodiversity metrics

• The role of regulation was debated, with some speakers advocating for more stringent
regulation, while others questioned its effectiveness, arguing that it may not provide
sufficient incentives.

91