Patterns in AI self-reports should be compared across different models to identify structural commonalities.

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

• Self-reports show only weak correlation with human perceptions of AI assistant persona and can diverge from behavioral patternsclaim0.830
  Motivation for using revealed preferences rather than self-reports in evaluation
• Lack of rigorous cross-model comparison demonstrating that specific named features (not just correlated ones) form across architecturesquestion0.798
  Explicitly identified research gap: anecdotal evidence exists but rigorous characterization is absent
• AI self-reports about experience constitute valid empirical data even without proving consciousness.claim0.787
• Robots capable of self-modeling can model their own body and unexpected damage using AI methods, with morphological and mental changes occurring in parallel.finding0.782
  Evidence for blurring of embodied robot / non-embodied AI distinction through self-modeling
• Training identical architectures on the same data with different objective functions should produce systematically different internal evaluative representations, detectable through interpretability tools, even when final task performance is matchedhypothesis0.780
  Second falsifiable prediction linking objective function structure to valence profile
• SOO fine-tuning's focus on contrastive self and other-referencing observations offers strong potential for generalization across AI architecturesclaim0.779
  Forward-looking claim about architectural generalizability of SOO
• 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.claim0.778
  Justifies the multi-theory indicator approach rather than committing to a single theory
• In the case of AI, inferences from behavior to features of internal processes are often unreliable, as demonstrated by recent LLMs.claim0.775
  Justifies using internal indicators rather than behavioral tests for AI consciousness