Measuring Goal-Directedness
This work addresses concerns about AI harm and philosophical aspects of agency by offering a tool to measure goal-directedness, though it appears incremental as it builds on existing maximum causal entropy frameworks.
The authors tackled the problem of quantifying goal-directedness in AI systems by introducing maximum entropy goal-directedness (MEG), a formal measure for causal models and Markov decision processes, and provided algorithms for its computation with small-scale experimental demonstrations.
We define maximum entropy goal-directedness (MEG), a formal measure of goal-directedness in causal models and Markov decision processes, and give algorithms for computing it. Measuring goal-directedness is important, as it is a critical element of many concerns about harm from AI. It is also of philosophical interest, as goal-directedness is a key aspect of agency. MEG is based on an adaptation of the maximum causal entropy framework used in inverse reinforcement learning. It can measure goal-directedness with respect to a known utility function, a hypothesis class of utility functions, or a set of random variables. We prove that MEG satisfies several desiderata and demonstrate our algorithms with small-scale experiments.