Undesirable Capabilities

Sub-category

Risk Domain

7AI System Safety, Failures & Limitations

7.6Multi-agent risks

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.

"Undesirable Capabilities. As agents interact, they iteratively exploit each other’s weaknesses, forc- ing them to address these weaknesses and gain new capabilities. This co-adaptation between agents can quickly lead to emergent self-supervised autocurricula (where agents create their own challenges, driving open-ended skill acquisition through interaction), generating agents with ever-more sophisticated strate- gies in order to out-compete each other (Leibo et al., 2019). This effect is so powerful that harnessing it has been critical to the success of superhuman systems, such as the use of self-play in algorithms like AlphaGo (Silver et al., 2016). However, as AI systems are released into the wild, it becomes possible for this effect to run rampant, producing agents with greater and greater capabilities for ends we do not understand"(p. 28)

Entity— Who or what caused the harm

Human

Due to a decision or action made by humans

AI system

Due to a decision or action made by an AI system

Other

Due to some other reason or is ambiguous

Intent— Whether the harm was intentional or accidental

Intentional

Due to an expected outcome from pursuing a goal

Unintentional

Due to an unexpected outcome from pursuing a goal

Other

Without clearly specifying the intentionality

Timing— Whether the risk is pre- or post-deployment

Pre-deployment

Occurring before the AI is deployed

Post-deployment

Occurring after the AI model has been trained and deployed

Other

Without a clearly specified time of occurrence

Supporting Evidence (1)

"For example, Baker et al. (2019) showed that even a simple game of hide and seek can lead to sophisticated tool use and coordination by MARL agents. In another case, researchers observed the emergence of manipulative communication, where an agent in an mixed-motive setting learns to use a shared communication channel to manipulate others (Blumenkamp & Prorok, 2021). Worse, this emer- gent complexity from co-adaptation could be open-ended and thus fundamentally unpredictable (Hughes et al., 2024)."(p. 29)

Part of Selection Pressures