What models can know about their own states, and when their reports are evidence.
Open questions
- How well can LLMs introspect
- Is there a global workspace
- Do models represent goals and preferences
- When self-report counts as welfare evidence
Papers
- Chalmers 2023 - Could a large language model be conscious
- Chalmers 2025 - Propositional interpretability
- Lindsey 2025 - Emergent introspective awareness
- Betley et al 2025 - Tell me about yourself
- Binder et al 2025 - Looking inward
- Comsa Shanahan 2025 - Introspection in LLMs
- Plunkett et al 2025 - Self-interpretability
- Kadavath et al 2022 - Language models mostly know what they know
- Gurnee Sofroniew et al 2026 - A global workspace in language models
- Dehaene Naccache 2026 - Does Claude possess a conscious global workspace
- Goldstein Kirk-Giannini 2024 - A case for AI consciousness
- Butlin Long et al 2023 - Consciousness in AI
- Perez Long 2023 - Self-reports and moral status
- Berg de Lucena Rosenblatt 2025 - Self-referential processing
- Chua Betley Marks Evans 2026 - The consciousness cluster
- Small LLMs 2026 - No reliable evidence of self-reported sentience
People
Placed by the editor, v1 — corrections welcome.