<span style="color: windowtext; mso-bidi-font-weight: bold;">Using Blockchain Ledgers to Record the AI Decisions in IoT<span style="color: windowtext;">

The integration of artificial intelligence (AI) into Internet of Things (IoT) systems has outpaced the development of mechanisms to explain and audit automated decisions, creating a transparency gap. This paper addresses the research problem of establishing immutable audit trails for AI-driven IoT decisions to enhance trust, accountability, and regulatory compliance. We propose a blockchain-based framework that logs each AI inference and its provenance data (inputs, model parameters, and outputs) on a tamper-proof distributed ledger, ensuring every decision is traceable and auditable. The technical method- ology centers on a permissioned blockchain ledger deployed alongside IoT infrastructure. IoT devices and edge nodes commit decision records via smart contracts, producing an im- mutable, timestamped log resistant to manipulation. This approach leverages blockchain&rsquo;s decentralization and cryptographic integrity to guarantee non-repudiation and data integrity. We detail how the system design balances transparency with privacy (e.g. hashing personal data) to remain compliant with data protection norms. The solution aligns closely with emerging regulatory frameworks such as the EU AI Act&rsquo;s mandate for automated decision logs and traceability, and GDPR&rsquo;s accountability and transparency requirements (e.g. maintaining audit logs of AI decisions for explainability). We demonstrate the frame- work&rsquo;s applicability across domains: healthcare IoT, to log diagnostic AI recommendations for accountability; and industrial IoT, to track autonomous control actions - showing that our approach generalizes to diverse high-stakes environments. The paper&rsquo;s contributions include a novel architecture for AI decision provenance in IoT, a detailed implementation on a blockchain ledger to securely record AI decision-making processes, and an evaluation of its performance and compliance benefits. By providing a reliable, immutable audit trail for AI in IoT, this work enhances transparency and trust in autonomous systems and offers a timely solution for auditable AI in an era of increasing regulatory scrutiny.