Trust, ethical exposure, and AI adoption intention among accounting professionals: Incremental explanatory evidence beyond UTAUT

Research on artificial intelligence (AI) adoption in accounting and auditing increasingly considers trust, ethics, and organizational conditions, but less often asks whether such appraisals add explanatory value after a conventional technology-acceptance baseline is already entered. This study conducts a composite-score nested-model comparison using 362 accounting professionals in Saudi Arabia, of whom 285 (78.7%) are external auditors. The UTAUT-aligned baseline contains performance expectancy (ETB), effort expectancy (EOP), social influence (PSE), and facilitating conditions (ERC); the added block contains perceived trust in AI (CTR) and ethical exposure appraisal (EEA). The baseline explains 25.2% of adoption-intention variance and the extended model explains 45.0% (ΔR 2 = .198; HC3 Wald χ 2 (2) =76.68, p < .001; f 2 = .36; 95% CI for ΔR 2 [.126, .283] from a 5,000-resample bootstrap). The increment remains positive after removal of the EEA item overlapping with willingness to use AI, alternative ETB/EOP baselines, demographic and occupational controls, conservative repeated-pattern rules, both chronological waves, and external-auditor subgroup analysis. CTR is positively associated with adoption intention, whereas EEA is also positive, contrary to its negative directional hypotheses; these individual coefficients are treated as secondary descriptive associations. Measurement evidence limits stronger interpretation: ETB and EOP are nearly indistinguishable (HTMT = .977), ordinal dimensionality analyses recover fewer dimensions than specified, and reliability deteriorates sharply in the later recruitment wave. The conclusion supported by the available evidence is therefore sample-bounded and model-comparative: the added score block contributes incremental in-sample explanatory information, but the study does not validate a new latent extension of UTAUT, establish stable mechanisms or causality, or measure actual AI use or audit quality.