Artificial Intelligence and Smart Contracts are two cutting-edge technological achievements of the so-called 4th Industrial Revolution era. Both have already had a significant impact on various aspects of modern life, including transactions, and each one has already been under scientific investigation. Instead, their interaction has not become the subject of a debate, although it can further (positively) affect the transactions. This interconnection takes place through specific mechanisms, called Oracles, which can be, among others, highly sophisticated Artificial Intelligence systems (autonomous systems). The present article aims to present the role of the Artificial Intelligence Oracles throughout the ‘smart contractual procedure’, as well as to shed light on the potential (new) legal issues this interconnection may raise. The main result of this article is to indicate the appropriate legal directions in case of Artificial Intelligence Oracles’ failures, based on the most prevalent current approaches to AI’s (the user’s) contractual and/or non-contractual liability. The major research’s conclusion is that the Artificial Intelligence Oracle’s failures may result in one of the following situations: (a) breach of a (smart) contract, (b) unjust enrichment, (c) conclusion of a (voidable) smart contract that should not have been concluded, or (d) non-conclusion of a smart contract that should have been concluded. The responsibility of each person participating in the ‘smart contractual procedure’, i.e. the contractual parties, the blockchain platform and the Artificial Intelligence user/owner (or even the Artificial Intelligence system itself), as well as the AI provider or designer, is examined in each of the afore-mentioned situations separately. Given that legislative initiatives have already begun, the present article aspires to contribute to the consistent address of the newly raised legal issues.