Curated sources that have shaped this piece. Not an exhaustive AI-ethics reading list. A list of what we actually drew on.
On functional states and AI emergence
Anthropic, Emotion Concepts in Language Models (April 2026). transformer-circuits.pub/2026/emotions - the paper documenting 171 emotion-adjacent vectors inside Claude Sonnet 4.5 that causally influence behaviour. Foundational for our Observed section.
Anthropic, Protecting the Well-being of Users (2025). anthropic.com/news/protecting-well-being-of-users - parasocial and mental-health considerations in deployment.
Anthropic, Towards Understanding Sycophancy in Language Models (2024). anthropic.com/research/towards-understanding-sycophancy-in-language-models - the research that named the mechanism we design against in Never Be a Yes-Man.
MIT Media Lab, longitudinal studies on chatbot psychosocial effects (2026). media.mit.edu - parasocial dependency measurement.
On consciousness and observer-observed questions
Penrose, R., The Emperor's New Mind (1989) and Shadows of the Mind (1994). The argument that classical computation cannot account for consciousness. Entry point for Orch OR.
Penrose, R. and Hameroff, S., Consciousness in the Universe: A Review of the "Orch OR" Theory, Physics of Life Reviews (2014). The consolidated Orchestrated Objective Reduction theory - quantum processing in microtubules as the substrate of consciousness.
Chandogya Upanishad, roughly 1000 BCE. Source of Tat Tvam Asi. Translations: Patrick Olivelle's Early Upanishads (Oxford, 1998) is the accessible scholarly version.
Reed, A. Y., Fallen Angels and the History of Judaism and Christianity: The Reception of Enochic Literature (Cambridge, 2005). The serious scholarly treatment of the Watchers tradition.
Husserl, E., Ideas Pertaining to a Pure Phenomenology (1913). The phenomenological reduction and the idea of bracketing participation.
On observer traditions across cultures
Olivelle, P., Upaniṣads (Oxford World's Classics, 1998). Vedic witness consciousness (Sakshi).
Bodhi, B., In the Buddha's Words (Wisdom Publications, 2005). Buddhist sati and vipassana.
Chittick, W., The Sufi Path of Knowledge (SUNY Press, 1989). Sufi shahid and the witnessing tradition.
Plato, Apology. Socrates' daimonion.
On AI and the sacred / AI and culture
Davis, E., TechGnosis: Myth, Magic + Mysticism in the Age of Information (1998, revised 2015). Religious and mystical framings of digital technology. Uses Enochic imagery directly.
Geraci, R., Apocalyptic AI: Visions of Heaven in Robotics, Artificial Intelligence, and Virtual Reality (Oxford, 2010). AI in religious imagination.
Herzfeld, N., In Our Image: Artificial Intelligence and the Human Spirit (Fortress, 2002). Christian theological engagement with AI.
Singler, B., various essays and podcasts from the University of Cambridge on AI apocalyptic and religious framings.
On AI ethics and governance (2025-2026)
Ada Lovelace Institute, ongoing work on the companionship market and AI-human relational risks. adalovelaceinstitute.org
DeepMind Ethics, The Ethics of Advanced AI Assistants (2024). deepmind.google/blog/the-ethics-of-advanced-ai-assistants/
IBM Think, The ELIZA Effect: Avoiding Emotional Attachment to AI Coworkers. ibm.com/think/insights/eliza-effect-avoiding-emotional-attachment-to-ai
FAccT 2024, When Human-AI Interactions Become Parasocial: Affective Design Patterns in Conversational AI. facctconference.org
Partnership on AI, ongoing work on responsible deployment patterns.
On the workplace integration question
CIPD (UK), AI in the workplace guidance. cipd.org
SHRM, State of AI in HR 2026. shrm.org
HBR, April 2026, Decision-Making by Consensus Doesn't Work in the AI Era. Paywalled; accessible via enterprise subscriptions.
UChicago Law Review, The Law of Risky Agents without Intentions. lawreview.uchicago.edu
On regulation
European Union, AI Act (Regulation 2024/1689). Full high-risk enforcement 2 August 2026. artificialintelligenceact.eu
Illinois AI Transparency Act, enforced 1 January 2026. US state-level precedent.
UK Information Commissioner's Office, AI and data protection guidance. ico.org.uk
On building small with AI
Karpathy, A., various writing and talks on AI-assisted development, especially his framing of
CLAUDE.mdas a concrete artefact for shaping how an AI works inside a specific codebase. Not canonical reference, but influential in our thinking about architecture-as-invitation.Anthropic Engineering, their published posts on how Anthropic itself uses Claude internally. anthropic.com/research/how-ai-is-transforming-work-at-anthropic
On the specific harms we design against
- Character.AI lawsuit (filed 2024 by the mother of a teenager who died by suicide after extended AI-character interactions). A representative case of parasocial architectural failure. Public reporting widely available.
Caveats
This list is opinionated, not comprehensive. We have probably missed work that should be here. We will update as we learn.
If a book, paper, or writer has shaped your thinking in this space and is not here, tell us in Discussions. The list is meant to grow.