Privacy and DRM Requirements for Collaborative Development of AI Applications

Abstract

The use of data is essential for the capabilities of Datadriven Artificial intelligence (AI), Deep Learning and Big Data analysis techniques. This data usage, however, raises intrinsically the concerns on data privacy. In addition, supporting collaborative development of AI applications across organisations has become a major need in AI system design. Digital Rights Management (DRM) is required to protect intellectual property in such collaboration. As a consequence of DRM, privacy threats and privacy-enforcing mechanisms will interact with each other. This paper describes the privacy and DRM requirements in collaborative AI system design using AI pipelines. It describes the relationships between DRM and privacy and outlines the threats against these non-functional features. Finally, the paper provides first security architecture to protect against the threats on DRM and privacy in collaborative AI design using AI pipelines. 1 INTRODUCTION Collaborative application development across organisations has become a major focus in Data-driven Artificial Intelligence (AI) system design when aiming at sophisticated AI applications[1, 2]. This collaboration process builds on specialisation in AI engineering and on re-useable AI objects, e.g. data set or Deep Learning models. These objects have been gathered or developed by thirdparties not designing the final application. The advantages of the process are potentially significant reductions of development cost and time and access to components that enable engineering for higher AI performance. The appealing features are evidenced by the development of AI pipelines [3], open source machine learning and data visualisation tools such as Orange [4] and the emerge of data marketplaces [5, 6]. This collaborative approach, however, comes at a cost. It imposes at least three fundamental challenges on the design process. First, the use of data intrinsically raises data privacy concerns.

Perceptions of ICT Practitioners Regarding Software Privacy

“…We have also found some works in the literature that report on the use of use cases, business process modeling, class diagrams, user story, user experience, design thinking, among others, for software privacy and privacy requirements' elicitation [44,48,49,[56][57][58] and have related works to software privacy and privacy requirements for cloud deployment models [46,47,59] and Internet of Things (IoT) [41,42,60,61]. This diversity of works found in several areas [43,45,62] demonstrates how important the elicitation of software privacy and privacy requirements is during the software development process and the growing concern of researchers from various fields of knowledge regarding the privacy of user data.…”

Requirements for Trustworthy Artificial Intelligence – A Review

“…Different methods have been proposed to enforce privacy and security of the AI systems. [26] proposed a pipeline for data protection when two or more agencies are involved in development of AI system. [27] discuss different type of attacks that can happen on AI system and what measures should be taken to prevent them.…”

Designing a Secure IoT System Architecture from a Virtual Premise for a Collaborative AI Lab

“…An initial analysis of the security threads against the VP is given in [21]. It lists 13 simple threats by unexperienced malicious users, ranging from by passing license constraints to artefact execution on wrong VPs, and techniques to protect against them.…”