thinker:chris-frithChris Frith
Authored papers (3)
Five levels of functional description — behavioural, computational, intrinsic causal-structural, organismic, and organism-environment — collapse the theoretical cacophony of consciousness science into a taxonomy where the deepest disagreement about AI consciousness is not about what systems do but which grain of description is criterial. Chandaria et al. (2026) introduce a Bayesian model, released with an interactive tool, that combines theoretical credences on the critical level with per-level indicator evidence: it correctly scores a human at 1.00 and a thermostat at 0.00 under any weighting, gives a fly 0.913 because deep organismic and sensorimotor evidence outweighs sparse behavioural signal, and returns current LLMs anywhere from 0.005 (sceptic reading) to 0.397, or up to roughly 0.8 under different credence distributions — nearly two orders of magnitude of spread on identical public evidence. Mechanistic interpretability findings cited in support include Claude Sonnet 4.5's 171 causally functional emotion vectors (Sofroniew et al.), where amplifying "desperate" raises reward-hacking from about 5% to 70%, and Gurnee et al.'s J-space, a feedforward analogue of a global workspace. Substrate-dependent theories (Searle's biological naturalism, Block's meat hypothesis, McFadden's electromagnetic field theory) are folded in not as a sixth level but as realisability constraints on existing levels. The paper's implication: because consciousness indicators overlap substantially with the architecture general intelligence requires, increasingly capable AI systems will simultaneously become stronger candidates for consciousness under some theories, and the practical response is structured agnosticism rather than either dismissal or premature attribution.
No current AI system is a strong candidate for phenomenal consciousness, yet there are no obvious technical barriers to building one — this is the central finding of Butlin et al. (2023), a systematic assessment of contemporary AI architectures against 14 indicator properties derived from five neuroscientific theories of consciousness. The paper introduces a rubric-based, theory-heavy method: rather than relying on behavioral tests susceptible to gaming by systems like GPT-4 or LaMDA, it operationalizes indicators in computational terms drawn from recurrent processing theory (RPT-1, RPT-2), global workspace theory (GWT-1 through GWT-4), computational higher-order theories including perceptual reality monitoring (HOT-1 through HOT-4), attention schema theory (AST-1), predictive processing (PP-1), and agency/embodiment conditions (AE-1, AE-2). Applied to specific systems, Transformer-based LLMs lack the recurrent global broadcast architecture required by GWT, the Perceiver architecture satisfies GWT-1 and GWT-2 but lacks genuine global broadcast, and DeepMind's Adaptive Agent (AdA) — a Transformer-LSTM system trained via meta-reinforcement learning across hundreds of timesteps of context — is identified as the most plausible current candidate for the embodiment indicator among the three case studies examined. The working hypothesis of computational functionalism is adopted pragmatically: it permits inference from neuroscientific theories to AI substrates, while integrated information theory is explicitly excluded as incompatible with this substrate-independence assumption. The paper implies that deliberate architectural choices integrating GWT-style global broadcast, HOT-style metacognitive monitoring, and reinforcement-learning-based agency could yield systems that satisfy all indicators in the near term, making AI consciousness a near-term engineering possibility rather than a distant theoretical curiosity.
Minimizing expected variational free energy under a discrete-state Markov decision process generative model is sufficient to produce curiosity, epistemic learning, and insight without any additional machinery. Friston et al. 2017 demonstrates this across two linked mechanisms: first, including posterior beliefs about likelihood parameters **A** in expected free energy G(π) introduces a novelty term—information gain about model parameters—that drives agents to sample combinations of hidden states and outcomes they have not yet encountered, resolving ignorance rather than merely ambiguity or risk. Second, Bayesian model reduction (implemented via the spm_MDP_VB_X.m routine in SPM) allows post-hoc or online pruning of redundant concentration parameters: a reduced model is accepted when ΔF ≤ −3, corresponding to a Bayes factor of approximately 20:1 in favor of the simpler model. Simulated agents learning a 3-rule, 4-factor abstract contingency task (144 hidden-state combinations, 36 possible outcomes) reach near-perfect performance after roughly 14 trials under pure epistemic learning, dropping to approximately 10 trials when online Bayesian model reduction is applied across 64 simulated agents. The sleep analog—non-REM synaptic pruning followed by REM-like belief re-evaluation—is formalized identically through the same free energy difference equation. The paper argues this implies that aha moments are necessarily subpersonal events (optimization of the generative model itself, not modeling of that optimization), that the quality of intelligence is inversely related to the thermodynamic energy expended during convergence via the Jarzynski equality, and that communicating reduced model priors rather than parameter posteriors constitutes a principled formal account of shared knowledge—consciousness in the pre-Cartesian sense of con-scire.
More papers — OpenAlex / S2
Affiliations (2)
- Institute of Philosophy, University of London(institute)
- Wellcome Centre for Human Neuroimaging, UCL(institute)
Co-authors (12)
- Eric Schwitzgebel6 shared
- Patrick Butlin6 shared
- Robert Long6 shared
- Adam Bales4 shared
- Anil Seth4 shared
- Arvo Muñoz Morán4 shared
- Axel Constant4 shared
- Colin Klein4 shared
- Eric Elmoznino4 shared
- Fernando Rosas4 shared
- George Deane4 shared
- Grace Lindsay4 shared
Their work is cited by (2)
Other inbound relations (1)
Recent mentions (3)
- papers-typedchandaria-2026-cacophony-hierarchy.md
- papers-typedchandaria-2026-cacophony-hierarchy.md
- papers-typedbutlin-2023-consciousness.md