Visibility into AI Agents

TL;DR

This paper explores methods to improve transparency and accountability of AI agents through identifiers, monitoring, and logging, analyzing their implementation, privacy implications, and impact on power dynamics.

Contribution

It provides a comprehensive assessment of visibility measures for AI agents, detailing their implementation, applicability, and societal implications.

Findings

Visibility measures vary in intrusiveness and informativeness.

Application of measures depends on deployment context and actors involved.

Enhanced visibility can support governance but raises privacy and power concerns.

Abstract

Increased delegation of commercial, scientific, governmental, and personal activities to AI agents -- systems capable of pursuing complex goals with limited supervision -- may exacerbate existing societal risks and introduce new risks. Understanding and mitigating these risks involves critically evaluating existing governance structures, revising and adapting these structures where needed, and ensuring accountability of key stakeholders. Information about where, why, how, and by whom certain AI agents are used, which we refer to as visibility, is critical to these objectives. In this paper, we assess three categories of measures to increase visibility into AI agents: agent identifiers, real-time monitoring, and activity logging. For each, we outline potential implementations that vary in intrusiveness and informativeness. We analyze how the measures apply across a spectrum of centralized through decentralized deployment contexts, accounting for various actors in the supply chain including hardware and software service providers. Finally, we discuss the implications of our measures for privacy and concentration of power. Further work into understanding the measures and mitigating their negative impacts can help to build a foundation for the governance of AI agents.