AI–Human Interaction — literature map

Index of the assembled secondary literature for the program. Organized by the twelve-branch HAI field map (see roadmap.md §I), with frontier-zone annotations. Status reflects what is in sources/ as of 2026-05-03.

Status legend:

• ✅ — review present in sources/topic-reviews/
• 🔁 — multi-LLM synthesis present in sources/syntheses/
• 🟡 — queued for next pass
• ⬜ — gap, not yet attempted

---

I. The twelve HAI branches

1. Interface and interaction design

Conversational UX, prompt design, multimodal I/O, latency, turn-taking. Foundational lineage: Nielsen, Norman, extended into the LLM era.

• ✅ Conversational AI interaction design after 2021 — sources/topic-reviews/conversational-ai-interaction-design-after-2021.md

2. Trust and reliance calibration

Automation bias, algorithmic aversion, the aversion-to-appreciation arc, trust-recovery dynamics. Anchors: Lee & See (2004), Parasuraman, Logg.

• ✅ Trust calibration and reliance under varying AI reliability in software work — sources/topic-reviews/trust-calibration-and-reliance-under-varying-ai-reliability-in-software-work.md

3. Explainability and interpretability (XAI)

LIME, SHAP, counterfactuals, mechanistic interpretability. Tension: technical interpretability vs. human-meaningful explanation.

• ✅ Explainability and interpretability for AI code generation — sources/topic-reviews/explainability-and-interpretability-for-ai-code-generation.md

4. Human-AI collaboration and teaming

Complementarity, division of labor, handoff dynamics, "centaur" workflows. Anchors: Shneiderman's HCAI framework.

• ✅ Organizational and sociotechnical embedding of AI coding agents — sources/topic-reviews/organizational-and-sociotechnical-embedding-of-ai-coding-agents.md
• ✅ Distributed cognition for AI-augmented professional work — sources/topic-reviews/distributed-cognition-for-ai-augmented-professional-work.md
• 🔁 Stigmergic coordination — sources/syntheses/stigmergic-coordination.md
• 🔁 Coordination-failure evidence in multi-agent LLM systems — sources/syntheses/coordination-failure-multi-agent-llm.md

5. Alignment and value learning

RLHF, constitutional methods, preference learning, debate, recursive reward modeling. Intersects moral philosophy and social choice theory.

• ⬜ Not yet drafted (lower priority for this program — alignment work is well-funded elsewhere; this program's contribution is downstream of alignment).

6. Social and parasocial dynamics

Anthropomorphism (Epley, Waytz, Cacioppo), CASA paradigm (Reeves & Nass), companionship AI, parasocial attachment.

• ✅ Parasocial dynamics, companion AI, and emotional engagement in extended human-AI interaction — sources/topic-reviews/parasocial-dynamics-companion-ai-and-emotional-engagement-in-extended-human-ai-interaction.md

7. Cognitive effects of AI use

Cognitive offloading, the "Google effect" extended to LLMs, deskilling, learning effects, metacognitive calibration.

• ✅ Longitudinal cognitive effects and skill change in AI-assisted programming — sources/topic-reviews/longitudinal-cognitive-effects-and-skill-change-in-ai-assisted-programming.md
• ✅ Cognitive apprenticeship, ZPD, and AI as a more-knowledgeable other — sources/topic-reviews/cognitive-apprenticeship-zpd-and-ai-as-a-more-knowledgeable-other.md
• 🔁 Ironies of automation in AI coding-agent supervision — sources/syntheses/ironies-of-automation-coding-agent-supervision.md
• 🔁 Risk compensation outside automotive safety — sources/syntheses/risk-compensation-outside-automotive-safety.md

8. Affective computing

Emotion recognition, emotional expression by AI, emotion regulation. Anchors: Picard.

• ⬜ Not yet drafted (Vela's emotion-corpus expansion at vela/docs/research/emotion-corpus-expansion/ overlaps; cross-link rather than duplicate).

9. Fairness, accountability, transparency, ethics (FAccT)

Bias auditing, fairness criteria, governance, contestability, harm taxonomies. Strong intersections with critical algorithm studies.

• ✅ Feminist epistemology, standpoint theory, and AI knowledge production — sources/topic-reviews/feminist-epistemology-standpoint-theory-and-ai-knowledge-production.md
• ✅ Political philosophy and democratic theory of AI after 2018 — sources/topic-reviews/political-philosophy-and-democratic-theory-of-ai-after-2018.md

10. Embodied and situated interaction

Social robotics, AR/VR agents, ambient AI. Anchors: Breazeal, presence and proxemics with non-human agents.

• ⬜ Not yet drafted (lower priority — not the program's near-term empirical surface).

11. Creativity and co-creation

AI as creative collaborator. Stylistic-homogenization debate, ideation vs. execution support.

• ✅ Creativity, ideation, and co-creation with AI — sources/topic-reviews/creativity-ideation-and-co-creation-with-ai.md
• ✅ Aesthetics of AI-generated art and AI-augmented practice — sources/topic-reviews/aesthetics-of-ai-generated-art-and-ai-augmented-practice.md

12. Persuasion, influence, and information ecosystem effects

AI's effects on beliefs, decisions, voting, purchasing; downstream collective epistemology.

• ✅ Persuasion, belief change, and AI's epistemic effects on users — sources/topic-reviews/persuasion-belief-change-and-ais-epistemic-effects-on-users.md

---

II. Frontier zones (across branches)

Longitudinal interaction

• ✅ Longitudinal cognitive effects and skill change in AI-assisted programming (cross-listed under Branch 7)

Multi-party configurations

• 🔁 Coordination-failure in multi-agent LLM systems (cross-listed under Branch 4)
• 🔁 Stigmergic coordination (cross-listed under Branch 4)

Long-context emergence

• ✅ Long-context emergence in conversational AI — sources/topic-reviews/long-context-emergence.md

Calibration of personalization

• ✅ Calibration of personalization in conversational AI — sources/topic-reviews/calibration-of-personalization.md

Cross-cultural and multilingual interaction

• ✅ Indigenous and non-Western relational ontologies in AI ethics and human-AI interaction — sources/topic-reviews/indigenous-and-non-western-relational-ontologies-in-ai-ethics-and-human-ai-interaction.md
• ✅ Communication accommodation theory in human-AI interaction — sources/topic-reviews/communication-accommodation-theory-in-human-ai-interaction.md

Sociotechnical embedding

• ✅ Organizational and sociotechnical embedding of AI coding agents (cross-listed under Branch 4)
• ✅ Institutional economics of AI and the reorganization of work and knowledge production — sources/topic-reviews/institutional-economics-of-ai-and-the-reorganization-of-work-and-knowledge-production.md

Failure modes and recovery

• ✅ Rupture and repair in human-AI coding interaction — sources/topic-reviews/rupture-and-repair-in-human-ai-coding-interaction.md
• ✅ Working alliance and rupture-repair from psychotherapy to human-AI interaction — sources/topic-reviews/working-alliance-and-rupture-repair-from-psychotherapy-to-human-ai-interaction.md

---

III. Adjacent fields imported (Roadmap §III)

These are the bodies of theory the HAI mainstream has under-engaged. The Penwright Research Program (Track A) draws materially from this set; Track B is built around bringing it forward.

• ✅ Companion-species and ethology of working partnerships — sources/topic-reviews/companion-species-and-ethology-of-working-partnerships.md
• ✅ Cognitive apprenticeship, ZPD, and AI as a more-knowledgeable other (cross-listed under Branch 7)
• ✅ Translation and interpreting studies for AI as a translator of domains and registers — sources/topic-reviews/translation-and-interpreting-studies-for-ai-as-a-translator-of-domains-and-registers.md
• ✅ Working alliance and rupture-repair from psychotherapy to human-AI interaction (cross-listed under Failure modes)
• ✅ Distributed cognition for AI-augmented professional work (cross-listed under Branch 4)
• ✅ Niche construction theory: feedback between organisms and environments — sources/topic-reviews/niche-construction-theory-feedback-between-organisms-and-environments.md
• ✅ Communication accommodation theory in human-AI interaction (cross-listed under cross-cultural)
• ✅ Improv theatre and ensemble performance for human-AI dialogue and collaboration — sources/topic-reviews/improv-theatre-and-ensemble-performance-for-human-ai-dialogue-and-collaboration.md
• ✅ Institutional economics of AI and the reorganization of work and knowledge production (cross-listed under sociotechnical)
• ✅ Phenomenology of skill and AI-augmented expert work — sources/topic-reviews/phenomenology-of-skill-and-ai-augmented-expert-work.md
• ✅ Phenomenology of attention in post-2010 theory and AI-augmented work — sources/topic-reviews/phenomenology-of-attention-in-post-2010-theory-and-ai-augmented-work.md
• 🔁 Transactive memory in human-AI teams — sources/syntheses/transactive-memory.md
• ✅ Ritual studies and non-human agents for long-term human-AI use — sources/topic-reviews/ritual-studies-and-non-human-agents-for-long-term-human-ai-use.md

---

IV. Deep intersections (Roadmap §IV)

Bodies of research where genuine theoretical integration is possible — not just analogy.

• ✅ Developmental psychology of children's interaction with AI — scoping note + annotated bibliography — sources/topic-reviews/developmental-psychology-children-ai.md (PA-003a; Option A scope; no-public-position discipline)
• ✅ Disability studies, crip theory, and AI as access infrastructure — sources/topic-reviews/disability-studies-crip-theory-and-ai-as-access-infrastructure.md
• ✅ Feminist epistemology, standpoint theory, and AI knowledge production (cross-listed under Branch 9)
• ✅ Indigenous and non-Western relational ontologies (cross-listed under cross-cultural)
• ✅ Psychoanalytic theory for human-AI working alliance measurement — sources/topic-reviews/psychoanalytic-theory-for-human-ai-working-alliance-measurement.md
• ✅ History of mediation technologies and AI — sources/topic-reviews/history-of-mediation-technologies-and-ai.md
• ✅ Memory, mourning, and digital remains — sources/topic-reviews/memory-mourning-and-digital-remains.md

---

V. Cross-LLM syntheses (the verifying layer)

These six documents take the secondary reviews above and re-check substantive claims against primary sources, with explicit claim-status tagging (A/B/C/D — see methodology.md §1.2).

• 🔁 Stigmergic Coordination — sources/syntheses/stigmergic-coordination.md
• 🔁 Coordination-Failure Evidence in Multi-Agent LLM Systems — sources/syntheses/coordination-failure-multi-agent-llm.md
• 🔁 Ironies of Automation in AI Coding-Agent Supervision — sources/syntheses/ironies-of-automation-coding-agent-supervision.md
• 🔁 Risk Compensation Outside Automotive Safety — sources/syntheses/risk-compensation-outside-automotive-safety.md
• 🔁 Transactive Memory — sources/syntheses/transactive-memory.md
• 🔁 CMU Faculty-Mapping Documents — sources/syntheses/cmu-faculty-mapping.md (program-design adjacency rather than topic synthesis)

---

VI. What's missing — honest gaps

• Long-context emergence — review shipped 2026-05-09 at sources/topic-reviews/long-context-emergence.md. Was the highest-leverage frontier gap; PA-001 closed it.
• Calibration of personalization — review shipped 2026-05-09 at sources/topic-reviews/calibration-of-personalization.md. PA-002 closed it.
• Developmental psychology of children with AI — scoping note shipped 2026-05-10 at sources/topic-reviews/developmental-psychology-children-ai.md. PA-003 Option A approved (narrow scoping note + bibliography); PA-003a closed. Program takes no public position on contested policy-adjacent claims per Option A discipline.
• Embodied / situated interaction — not yet attempted; lower priority for this program's near-term empirical surface.
• Alignment and value learning — explicitly out of scope; well-covered elsewhere.
• Affective computing — Vela's emotion-corpus-expansion/ covers adjacent ground; cross-link rather than duplicate.
• Conversation analysis and ethnomethodology — review shipped 2026-05-09 at sources/topic-reviews/conversation-analysis-ethnomethodology.md. PA-004 closed it.

---

Cross-references

• roadmap.md — the field-mapping document this map is indexed against
• program.md — the twelve-paper Penwright Research Program drawing from this literature
• methodology.md — how the reviews and syntheses were generated
• vela/docs/research/emotion-corpus-expansion/ — adjacent program at Vela; affective-computing overlap