Towards Sustainable Web Agents: A Plea for Transparency and Dedicated Metrics for Energy Consumption

2025-02-25

AI Summary

This paper addresses the lack of transparency and standardized energy-efficiency metrics for web agents in sustainability assessment. We propose, for the first time, a dedicated energy-efficiency evaluation paradigm tailored to such systems. Methodologically, we integrate empirical power consumption modeling, CO₂ emission estimation, and process-level energy attribution analysis to conduct横向 cross-agent efficiency comparisons under representative tool-calling scenarios. Results reveal that per-task energy consumption varies by several-fold across agents; architectural design choices significantly impact efficiency; and opaque model parameters and execution workflows severely hinder quantitative sustainability assessment. Our primary contributions are: (1) establishing the first methodological framework for green evaluation of web agents, (2) identifying critical energy bottlenecks—particularly in orchestration logic and API interaction patterns—and (3) advocating the integration of transparency and sustainability metrics into core evaluation criteria to advance reproducible, auditable green AI standards.

Technology Category

• Machine Learning: Efficient ML / Green AI
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Modeling other Agents

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Sustainability and climate impact of web technologies
• Economics, Online Markets and Human Computation: Sustainability of Web economics

Abstract

Improvements in the area of large language models have shifted towards the construction of models capable of using external tools and interpreting their outputs. These so-called web agents have the ability to interact autonomously with the internet. This allows them to become powerful daily assistants handling time-consuming, repetitive tasks while supporting users in their daily activities. While web agent research is thriving, the sustainability aspect of this research direction remains largely unexplored. We provide an initial exploration of the energy and CO2 cost associated with web agents. Our results show how different philosophies in web agent creation can severely impact the associated expended energy. We highlight lacking transparency regarding the disclosure of model parameters and processes used for some web agents as a limiting factor when estimating energy consumption. As such, our work advocates a change in thinking when evaluating web agents, warranting dedicated metrics for energy consumption and sustainability.

Problem

• Assessing energy consumption of web agents
• Exploring sustainability in web agent design
• Advocating for transparency in agent metrics

Innovation

• Web agents use external tools
• Advocate energy consumption metrics
• Highlight transparency in model parameters