Minimal Computational Preconditions for Subjective Perspective in Artificial Agents
This work addresses the challenge of creating measurable signatures of subjectivity in machine systems, which is incremental as it builds on existing phenomenological and computational frameworks.
The study tackled the problem of operationalizing subjective perspective in artificial agents by implementing a slowly evolving global latent state that modulates policy dynamics, resulting in direction-dependent hysteresis in a reward-free environment with regime shifts.
This study operationalizes subjective perspective in artificial agents by grounding it in a minimal, phenomenologically motivated internal structure. The perspective is implemented as a slowly evolving global latent state that modulates fast policy dynamics without being directly optimized for behavioral consequences. In a reward-free environment with regime shifts, this latent structure exhibits direction-dependent hysteresis, while policy-level behavior remains comparatively reactive. I argue that such hysteresis constitutes a measurable signature of perspective-like subjectivity in machine systems.