Skip to content

R0029/2026-03-27/Q001

Query: Has anyone proposed a mechanism, standard, or framework for attributing AI contributions to collaborative human-AI work? Look at academic institutions, publishers, professional organizations, and standards bodies for formal proposals, guidelines, or emerging norms.

BLUF: Multiple structured proposals exist — including the IBM AI Attribution Toolkit (May 2025), the AIA icon system (Avery et al., 2024), and CHI 2025 research on attribution perceptions — but no standards body has adopted an AI-specific attribution standard. The field is in an active proposal phase, pre-standardization. The existing CRediT taxonomy (NISO) covers human contributors only.

Answer: H3 (Emerging but pre-standard) · Confidence: Medium


Summary

Entity Description
Query Definition Question as received, clarified, ambiguities, sub-questions
Assessment Full analytical product
ACH Matrix Evidence × hypotheses diagnosticity analysis
Self-Audit ROBIS-adapted 4-domain process audit

Hypotheses

ID Statement Status
H1 Multiple formal frameworks exist Partially supported
H2 Only ad hoc disclosure norms exist Eliminated
H3 Emerging but pre-standard Supported

Key Frameworks Identified

Framework Proposer Year Type Status
AI Attribution Toolkit IBM Research (He, Weisz, Houde) 2025 Working questionnaire tool Experimental / "first pass"
AIA Icon System Avery, Abril & del Riego 2024 Visual icon system Law review proposal
CHI 2025 Research He et al. / IBM 2025 Empirical attribution research Published findings
CRediT Taxonomy NISO/CASRAI 2022 14-role contributor taxonomy Adopted standard (human-only)

Searches

ID Target Type Outcome
S01 Attribution frameworks and standards WebSearch 3 selected, 7 rejected
S02 IBM toolkit and CHI research WebSearch 1 selected, 9 rejected

Sources

Source Description Reliability Relevance Evidence
SRC01 He et al., CHI 2025 High High 1 extract
SRC02 IBM AI Attribution Toolkit Medium-High High 1 extract
SRC03 Avery, Abril & del Riego, JTIP High High 1 extract
SRC04 CRediT/NISO High Medium 1 extract

Revisit Triggers

  • NISO or ISO announces an AI-specific extension to CRediT
  • Any publisher adopts a structured attribution framework beyond binary disclosure
  • IBM toolkit or AIA system achieves measurable adoption (>100 institutional users)