GWT advocates should clarify whether the system built to implement all four GWT indicators is conscious, given that the indicator method implies a positive assessment.

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Satisfying the indicators would not mean that an AI system would definitely be conscious.claim0.802
  Caveat that indicators are not conclusive proof.
• Indicators used to assess AI systems for consciousness should focus on theories' central explanatory posits, abstracting away from implementation details that may differ across systems.claim0.791
  First of four guidelines for deriving indicators; prevents over-restriction to human-specific features
• AI systems which possess more of the indicator properties are more likely to be conscious.claim0.780
  Graded claim about the rubric.
• Indicators are merely intended to be credence-shifting; we do not need to be certain that a theory is correct for it to provide useful indicators.claim0.780
  Key epistemic stance of the Bayesian approach to indicators
• Because no one theory of consciousness is currently dominant, assessments of AI systems should draw on multiple theories and treat their conditions as indicators rather than definitive criteria.claim0.775
  Justifies the multi-theory indicator approach rather than committing to a single theory
• Perceiver architecture is some way from being a full implementation of GWT.claim0.768
  Case study conclusion.
• We hypothesize that interpretability methods could provide further evidence about indicators in particular systems or serve as the basis for distinct tests for consciousness.hypothesis0.768
  Forward-looking suggestion for how inner interpretability could extend the indicator method
• Indicators should avoid ambiguous terms and contested concepts, but without prematurely committing to precise specifications that anticipate all uncertainties.claim0.767
  Fourth guideline for deriving indicators; illustrated by algorithmic vs implementation-level recurrence distinction