EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Epistemic norms for AI safety and alignment research

Keivan Navaie

Artificial Intelligence Review · 07.08.2026 · 59 (12)

DOI: 10.1007/s10462-026-11666-8 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Artificial Intelligence Review, online erschienen am 07.08.2026. Verfasst von Keivan Navaie. Die Arbeit nennt 99 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Mainstream AI research emphasises capability growth and tolerates low failure rates when average-case performance is high. AI safety and alignment research has a different mission: to ensure that catastrophic failures never occur, under sparse evidence, adversarial dynamics, and fat-tailed risk. We argue that the two domains differ along two analytically independent axes — capability profile (demonstrating the absence of hazardous behaviours versus the presence of positive capabilities) and risk profile (bounding worst-case outcomes under fat-tailed uncertainty versus optimising average-case performance) — and that mainstream epistemic practices are inadequate on both. Building on a structured synthesis grounded in a preregistered bibliometric baseline, we identify five cross-cutting gap dimensions in current alignment research, including the near-absence of institutionalised independent verification. To address these gaps we propose ECAISA, an Epistemic Code for AI Safety and Alignment comprising eight principles, a three-level scoring rubric, a four-level disclosure ladder that reconciles transparency with information-hazard and commercial-confidentiality constraints, a tiered applicability scheme, an infohazard adjudication procedure, and seven anti-gaming mechanisms. ECAISA does not certify that any AI system is safe; it constrains how safety-relevant research claims are documented, checked, and relied upon, with auditability rather than certification as its governance target. A retrospective rubric audit (κ = 0.79) demonstrates instrument feasibility; a four-stage validation roadmap is proposed.

Zitieren

APA

Navaie, K. (2026). Epistemic norms for AI safety and alignment research. Artificial Intelligence Review, 59(12). https://doi.org/10.1007/s10462-026-11666-8

Vancouver

Navaie K. Epistemic norms for AI safety and alignment research. Artificial Intelligence Review. 2026;59(12). doi:10.1007/s10462-026-11666-8

BibTeX
@article{Navaie2026,
  author = {Navaie, Keivan},
  title = {Epistemic norms for AI safety and alignment research},
  journal = {Artificial Intelligence Review},
  year = {2026},
  volume = {59},
  number = {12},
  doi = {10.1007/s10462-026-11666-8}
}
Kostenlos starten →