APPRAISE: a Governance Framework for Innovation with Artificial Intelligence Systems

Abstract

As artificial intelligence (AI) systems increasingly impact society, the EU Artificial Intelligence Act (AIA) is the first legislative attempt to regulate AI systems. This paper proposes a governance framework for organizations innovating with AI systems. Building upon secondary research, the framework aims at driving a balance between four types of pressures that organizations, innovating with AI, experience, and thereby creating responsible value. These pressures encompass AI/technology, normative, value creation, and regulatory aspects. The framework is partially validated through primary research in two phases. In the first phase, a conceptual model is proposed that measures the extent to which organizational tasks result in AIA compliance, using elements from the AIA as mediators and strategic variables such as organization size, extent of outsourcing, and offshoring as moderators. 34 organizations in the Netherlands are surveyed to test the conceptual model. The average actual compliance score of the 34 participants is low, and most participants exaggerate their compliance. Organization size is found to have significant impact on AIA compliance. In phase 2, two case studies are conducted with the purpose of generating in-depth insights to validate the proposed framework. The case studies confirm the interplay of the four pressures on organizations innovating with AI, and furthermore substantiate the governance framework.