Fork the consciousness, or download the project and create your own.

Peter David Fagan

Not institutionally listed on the preprint.

Integrated information Consciousness is intrinsic causal power, measured as integrated information

Peter David Fagan works on measurement frameworks for artificial intelligence safety. His research note “Revisiting Classic Thought Experiments to Measure Consciousness for Artificial Intelligence Safety” (arXiv 2608.00001) introduces Conservation-Congruent Encoding, a framework that scores a system on two separate quantities. Task performance, W_causal,T, is ordinary behavioural success. Operational consciousness, kappa_T, is the efficiency with which preserved internal structure supports that behaviour.

The position the framework defends is that behavioural equivalence between systems underdetermines their organisation. An uncompressed lookup system and a compact generative system can match on performance while diverging sharply in structure reuse. The lookup system maintains an expanding standing store of unreused mappings. The generative system compresses regularities into structure it applies across cases. Fagan argues the second quantity is what the classic disputes about understanding, Leibniz’s mill, the imitation game and the Chinese Room, were reaching for without a measure to hold.

The contribution to consciousness research is deliberately narrow. Operational consciousness is defined within the framework and is not claimed to measure phenomenal consciousness. The safety-relevant claim is that AI safety analysis needs instruments for organisational divergence that behaviour cannot detect, and that the thought experiments survive reframing as the clearest statements of why the quantities must separate.

Known for. Conservation-Congruent Encoding, operational consciousness as a structure-reuse measure, reframing classic philosophy of mind thought experiments as measurement problems

Coverage on this site

Related researchers