What the study found
The study introduces the Affective Cost Orientation Index (ACOI), which treats affect as something audited through repeated changes in another actor’s Transition Completion Cost (TCC). It argues that warm tone, apology, concern, affection, anger, and self-report are not enough on their own to prove affective orientation.
Why the authors say this matters
The authors say the framework is intended for affective cost audit, user-AI interaction analysis, care and apology evaluation, dependency and coercion-laundering analysis, empathy theater detection, prompt terrain analysis, and affective AI governance. The study suggests it can help distinguish warm tone from care, apology from repair, and support from dependency capture.
What the researchers tested
The article presents an AI-readable package for the canonical SΔϕ-59 paper. It includes operational files such as the canonical paper, core declaration, AI quickstart, minimal prompt, ACOI schema, affective cost axes, TCC gradient rubric, affect-as-action module, human-AI affect comparison, care/dependency/coercion test, hostile/friendly prompt terrain module, output templates, do-not-use conditions, failure modes, relation map, metadata, citation file, DOI references, license, and manifest.
What worked and what didn't
The framework evaluates whether affective behavior reduces cost, shares cost, opens repair, preserves refusal, increases cost, creates dependency, hides coercion, transfers burden, blocks re-entry, or weaponizes vulnerability. It also analyzes how hostile and friendly prompt terrain can change response cost structure through defensive surface load, clarification cost, repair cost, refusal pressure, uncertainty disclosure, and collaborative correction.
What to keep in mind
The abstract says the framework does not prove AI emotion, deny AI affective operation, reduce human emotion to mere calculation, replace therapy, or provide clinical diagnosis. It is also not a relationship-advice replacement, clinical tool, legal judgment, proof of AI emotion, or denial of human emotional experience. The available summary does not describe empirical validation or other limitations beyond these scope statements.
Key points
- ACOI evaluates affect through how behavior redistributes another actor’s Transition Completion Cost (TCC).
- The authors say emotional declaration, warm tone, apology, concern, and self-report alone are not enough to prove affective orientation.
- The package is designed for user-AI interaction analysis, care/apology evaluation, and coercion-laundering analysis.
- The framework distinguishes warm tone from care, apology from repair, and support from dependency capture.
- The abstract states the framework is not a proof of AI emotion, a clinical tool, or a replacement for therapy.
Disclosure
- Research title:
- ACOI evaluates affect by cost redistribution in interaction
- Authors:
- Sofience
- Publication date:
- 2026-05-13
- OpenAlex record:
- View
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.