Masaya Ochiai

Artificial intelligence (AI) ethics has expanded rapidly through the global circulation of principle vocabularies such as transparency, fairness, and accountability. Yet lexical alignment can conceal how responsibility and legitimacy are reallocated when principles are translated into institutions and operations. This paper identifies a recurrent comparison-design failure in intra-Asian comparison: when Euro-American frameworks silently become the reference point, intra-regional comparison degenerates into deviation scoring and cannot jointly handle high internal plurality and cross-border interdependence. I conceptualize this failure as reference-point drift. To diagnose and reduce drift risk, I propose a minimal procedure called normative translation audit (NTA). NTA redesigns the unit of comparison from principle vocabularies to four branch points that determine practical governance: (1) operational definitions, (2) veto rights, (3) remedy lines, and (4) accountability assignment. NTA is not a pass/fail compliance audit or third-party assurance; it is a minimal procedure for fixing comparison units and tracing whether register updates actually changed decisions in specifications, operations, or remedy design. NTA requires three artifacts: a Branch Register, a Decision-Change Log, and a multi-anchor memo that documents anchor selection and exclusion so that external frameworks do not become the only baseline. A brief demonstration (Association of Southeast Asian Nations and Singapore) illustrates how branch-point comparison reveals divergences—especially around contestation and decision review—that vocabulary-based comparison tends to hide, while outlining a testable protocol for future empirical evaluation.

Large language models are increasing demand for high-quality training corpora, yet scholarly publishing lacks scalable and legitimacy-preserving mechanisms for governing the inclusion of research articles in training datasets. Back-catalogue agreements concluded without transparent governance can damage trust, while retroactive opt-in campaigns impose high transaction costs and uncertain uptake. This article proposes the ‘Third Option’: a post-acceptance, opt-in addendum in which authors authorize defined artificial intelligence (AI) training uses of the accepted article in exchange for an author-facing benefit. The addendum is designed to be choice-expanding and editorially independent, and to operate across publishing models, including diamond open access journals that do not charge article processing charges. Its value is not limited to copyright permission: it can bundle provenance metadata, rights hygiene for third-party content, dataset documentation, versioning and contractual assurances on permitted uses and auditability. The author defines withdrawal semantics as future-only exclusion from subsequent dataset releases, while acknowledging the practical limits of machine unlearning. This article formalizes five propositions, addresses key objections and outlines a methods-style evaluation protocol for empirical testing.

Large language models (LLMs) are increasingly used for drafting, translating, and editing academic prose. Recent debates and editorial policies often treat facts about how a text was produced as epistemically significant. This paper diagnoses the philosophical assumption behind such process-oriented norms: process essentialism, the view that the justificatory status of academic knowledge claims is constitutively dependent on their production process. I argue that strong forms of process essentialism are mistaken once we distinguish justificatory status from ethical questions of credit, accountability, and provenance. In publication contexts, what primarily matters is the quality of the public epistemic record—argumentative structure, evidential support, engagement with relevant work, and, where applicable, data, code, methods, and provenance claims—rather than the causal history of composition. I develop three objections: a genetic-fallacy argument, an argument from the unobservability and unenforceability of private process criteria, and an epistemic justice argument about exclusionary burdens. I conclude by showing how disclosure, contributorship, and provenance practices can be justified within an output-based framework without treating private process facts as constitutive of justification.