Key points are not available for this paper at this time.
AI has found widespread application across various sectors, including security, healthcare, finance, and national defense. However, alongside its transformative advancements, there has been an unfortunate trend of malicious exploitation of AI capabilities. Concurrently, the rapid evolution of cloud computing technology has introduced cloud-based AI systems. Regrettably, vulnerabilities inherent in cloud computing infrastructure also pose risks to the security of AI services. We observe that the integrity of training data is pivotal, as any compromise therein directly impacts the efficacy of AI systems. Against this backdrop, we assert the paramount importance of preserving data integrity within AI systems. To address this imperative, we propose a data integrity architecture guided by the National Institute of Standards and Technology (NIST) cyber security framework. Leveraging block chain technology and smart contracts emerges as a fitting solution to tackle integrity challenges, owing to their characteristics of shared and decentralized ledgers. Smart contracts facilitate automated policy enforcement, enable continuous monitoring of data integrity, and mitigate the risk of data tampering.
Bayani et al. (Fri,) studied this question.