The HATI² Framework Honest, Accurate, Transparent, Insightful A Dual-Axis Evaluation Framework for Human-AI Collaborative Intellectual Production — Version 1.0

ABSTRACT The HATI2 framework is a dual-axis evaluation methodology for assessing the quality, integrity, and scholarly contribution of human-AI collaborative intellectual work. It was developed through iterative application across the ten papers of the SymbioMind Bonding Relationships Series (2026) and is now formalised as a standalone tool available to the broader research community. The framework operates on two axes: Tier 1 (SIP Core) assesses Sovereign Intellectual Property integrity across five dimensions — Honesty, Accuracy, Transparency, Insight, and Impact — with a combined weight of 50 points. Tier 2 (Academic Rigor) assesses scholarly quality across four dimensions — Hermeneutic coherence, Analytical logic, Theoretical robustness, and Implementation guidance — with a combined weight of 50 points. The two tiers are evaluated independently and combined for a total score of 100. Minimum thresholds are specified for SIP inclusion (Tier 1 > 35), publication readiness (combined > 75), and flagship status (combined > 90). The framework addresses a gap in current scholarly evaluation infrastructure: existing peer review processes were designed for single-author or multi-author human work and do not adequately distinguish the specific integrity requirements, attribution standards, and quality dimensions relevant to sustained human-AI intellectual collaboration. HATI2 provides a structured, reproducible, and publicly documented alternative. Calibration data from the SymbioMind Bonding Relationships Series (Papers 1–10, March 2026) is provided as the framework’s validation corpus. Score distributions, dimension-level analysis, and inter-evaluator guidance are included. Keywords: evaluation framework, human-AI collaboration, scholarly integrity, dual-axis evaluation, SIP Commons, SymbioMind, HATI, peer review alternative, working paper evaluation, collaborative epistemology