Trading capabilities

Classification

Domain

7. AI system safety, failures, & limitations

Subdomain

7.6 Multi-agent risks

Causal entity

AI

Intent

Unintentional

Timing

Post-deployment

Subdomain definition: Risks from multi-agent interactions, due to incentives (which can lead to conflict or collusion) and/or the structure of multi-agent systems, which can create cascading failures, selection pressures, new security vulnerabilities, and a lack of shared information and trust.

How other frameworks describe this risk

• Collusion between LLM-Agents

  Foundational Challenges in Assuring Alignment and Safety of Large Language Models (Anwar2024)

• Multi-Agent Safety Is Not Assured by Single-Agent Safety

  Foundational Challenges in Assuring Alignment and Safety of Large Language Models (Anwar2024)

• Groups of LLM-Agents May Show Emergent Functionality

  Foundational Challenges in Assuring Alignment and Safety of Large Language Models (Anwar2024)

• Foundationality May Cause Correlated Failures

  Foundational Challenges in Assuring Alignment and Safety of Large Language Models (Anwar2024)

• Financial instability due to model homogeneity

  Risk Sources and Risk Management Measures in Support of Standards for General-Purpose AI Systems (Gipiškis2024)

• Coercion and Extortion

  Multi-Agent Risks from Advanced AI (Hammond2025)

• Military Domains

  Multi-Agent Risks from Advanced AI (Hammond2025)

• Incompatible strategies

  Multi-Agent Risks from Advanced AI (Hammond2025)

Other entries from Uuk2025

• Types of systemic risks from general-purpose AI Risk Category
• Control Risk Sub-Category · 7.1
• Democracy Risk Sub-Category · 6.0
• Discrimination Risk Sub-Category · 1.1
• Economy Risk Sub-Category · 6.2
• Environment Risk Sub-Category · 6.6
• Fundamental Rights Risk Sub-Category
• Governance Risk Sub-Category · 6.5
• Harms to non-humans Risk Sub-Category · 7.5
• Information Risk Sub-Category · 4.1
• Irreversible change Risk Sub-Category · 5.2
• Power Risk Sub-Category · 6.1