TY  - JOUR
T1  - Trust Architectures for AI-Enabled Drug Discovery: Transparency, Validation, Reproducibility, Accountability, and Regulatory Alignment
A1  - Saif Al-Hinai
A1  - Amal Al-Balushi
A1  - Mohammed Al-Maskari
JF  - International Journal of Pharmaceutical And Phytopharmacological Research
JO  - Int J Pharm Phytopharmacol Res
SN  - 2250-1029
Y1  - 2026
VL  - 16
IS  - 2
DO  - 10.51847/nI2B8SaJLi
SP  - 106
EP  - 122
N2  - Artificial intelligence is increasingly embedded in drug discovery through molecular representation, target identification, property prediction, virtual screening, candidate prioritization, and evidence interpretation. However, predictive performance alone cannot establish that an AI system is trustworthy, scientifically credible, or suitable for consequential decision support. This article proposes a conceptual trust architecture that organizes technical, evidentiary, organizational, and lifecycle requirements for AI-enabled drug discovery. The architecture treats transparency as a structured combination of data provenance, dataset documentation, model documentation, traceability, interpretability, explainability, and uncertainty reporting. Validation is organized through predefined claims, context-sensitive benchmarks, leakage controls, internal and external evaluation, and prospective or experimental assessment where relevant. Reproducibility is supported through documented workflows, version control, accessible computational resources where appropriate, and reconstructable model-development records. Accountability is addressed through human oversight, assigned governance roles, auditable decisions, bias and fairness assessment, risk management, and explicit escalation boundaries. Regulatory alignment is framed as the organization of evidence, monitoring, change control, and implementation safeguards rather than evidence of regulatory approval or legal sufficiency. The proposed architecture also incorporates lifecycle monitoring, drift detection, controlled model updating, and feedback governance. Its principal contribution is an integrated framework for connecting trustworthy-AI principles with the practical evidence and decision structures of drug discovery while maintaining clear boundaries between governance readiness, implementation readiness, clinical utility, and regulatory authorization.
UR  - https://eijppr.com/article/trust-architectures-for-ai-enabled-drug-discovery-transparency-validation-reproducibility-accoun-a1ph4ymhydso0ry
ER  -