From Paradox to Progress A Net Positive AI Energy Framework 2025
Page 19 of 38 · WEF_From_Paradox_to_Progress_A_Net_Positive_AI_Energy_Framework_2025.pdf
Shape demand wisely – key levers FIGURE 7
Use-based
pricing modelsDigital sobriety
campaignsModel selection
guidanceConsumer
dashboardsRegulatory
nudgesShape demand
wisely
Use case insights and takeaways:
Only 10% of use cases reflected shape demand
wisely, underscoring a gap or limited visibility into
ecosystem activities here. These examples aim to
influence AI consumption patterns and promote
selective, energy-aware applications. Without
greater attention to this driver, efficiency gains risk
being offset by continued AI demand growth.Strategic recommendations:
–Establish AI sustainability labels and energy
disclosures.
–Provide public dashboards with opt-in
transparency.
–Promote responsible defaults and user awareness.
Emerging shape demand wisely use case examples TABLE 4
Source: AI Energy Impact public use case database.37Use-based
pricing
modelsCloud provider pilots (tiered pricing): Ties cost to energy intensity, incentivizing efficient AI use and
moderating demand
Digital
sobriety
campaignsGlobal consulting firm (digital sobriety programme): This programme drove digital sustainability awareness
and audits to cut unnecessary compute and improve energy efficiency
Model
selection
guidanceLeboncoin (API traffic optimization): Detected 3.6 billion redundant API calls weekly, cutting error responses by 72%
and server replicas 67%
Consumer
dashboardsGlobal software company (AI energy score prototype): A dashboard displayed the energy impact of every query,
helping users select lower-energy workflows
Regulatory
nudgesGlobal media, entertainment and sports company (network efficiency programme): Regulator partnership
optimized network operations, reducing energy use by approximately 40% and meeting efficiency standards
From Paradox to Progress: A Net-Positive AI Energy Framework
19
Ask AI what this page says about a topic: