Matti Mäntymäki

Advancements in hardware and software have propelled machine learning (ML) solutions to become vital components of numerous information systems. This calls for research on the integration and evaluation of ML development practices within software companies. To investigate these issues, we conducted expert interviews with software and ML professionals. We structured the interviews around information systems development (ISD) models, which serve as conceptual frameworks that guide stakeholders throughout software projects. Using practice theory, we analyzed how software professionals perceive ML development within the context of ISD models and identified themes that characterize the transformative impact of ML development on these conceptual models. Our findings show that developer-driven conceptual models, such as DevOps and MLOps, have been embraced as common frameworks for developers and management to understand and guide the ML development processes. We observed ongoing shifts in predefined developer roles, wherein developers are increasingly adopting ML techniques and tools in their professional work. Overall, our findings underscore that ML technologies are becoming increasingly prominent in software projects across industries, and that the incorporation of ML development in ISD models is an ongoing, largely practice-driven, process.

Purpose: Following the surge of documents laying out organizations' ethical principles for their use of artificial intelligence (AI), there is a growing demand for translating ethical principles to practice through AI governance (AIG). AIG has emerged as a rapidly growing, yet fragmented, research area. This paper synthesizes the organizational AIG literature by outlining research themes and knowledge gaps as well as putting forward future agendas.Design/methodology/approach: The authors undertake a systematic literature review on AIG, addressing the current state of its conceptualization and suggesting future directions for AIG scholarship and practice. The review protocol was developed following recommended guidelines for systematic reviews and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).Findings: The results of the authors’ review confirmed the assumption that AIG is an emerging research topic with few explicit definitions. Moreover, the authors’ review identified four themes in the AIG literature: technology, stakeholders and context, regulation and processes. The central knowledge gaps revealed were the limited understanding of AIG implementation, lack of attention to the AIG context, uncertain effectiveness of ethical principles and regulation, and insufficient operationalization of AIG processes. To address these gaps, the authors present four future AIG agendas: technical, stakeholder and contextual, regulatory, and process. Going forward, the authors propose focused empirical research on organizational AIG processes, the establishment of an AI oversight unit and collaborative governance as a research approach.Research limitations/implications: To address the identified knowledge gaps, the authors present the following working definition of AIG: AI governance is a system of rules, practices and processes employed to ensure an organization's use of AI technologies aligns with its strategies, objectives, and values, complete with legal requirements, ethical principles and the requirements set by stakeholders. Going forward, the authors propose focused empirical research on organizational AIG processes, the establishment of an AI oversight unit and collaborative governance as a research approach.Practical implications: For practitioners, the authors highlight training and awareness, stakeholder management and the crucial role of organizational culture, including senior management commitment.Social implications: For society, the authors review elucidates the multitude of stakeholders involved in AI governance activities and complexities related to balancing the needs of different stakeholders.Originality/value: By delineating the AIG concept and the associated research themes, knowledge gaps and future agendas, the authors review builds a foundation for organizational AIG research, calling for broad contextual investigations and a deep understanding of AIG mechanisms. For practitioners, the authors highlight training and awareness, stakeholder management and the crucial role of organizational culture, including senior management commitment.

While there is widespread consensus that artificial intelligence (AI) needs to be governed owing to its rapid diffusion and societal implications, the current scholarly discussion on AI governance is dispersed across numerous disciplines and problem domains. This paper clarifies the situation by discerning two problem areas, metaphorically titled the “easy” and “hard” problems of AI governance, using a dialectic theory synthesis approach. The “easy problem” of AI governance concerns how organizations’ design, development, and use of AI systems align with laws, values, and norms stemming from legislation, ethics guidelines, and the surrounding society. Organizations can provisionally solve the “easy problem” by implementing appropriate organizational mechanisms to govern data, algorithms, and algorithmic systems. The “hard problem” of AI governance concerns AI as a general-purpose technology that transforms organizations and societies. Rather than a matter to be resolved, the “hard problem” is a sensemaking process regarding socio-technical change. Partial solutions to the “hard problem” may open unforeseen issues. While societies should not lose track of the “hard problem” of AI governance, there is significant value in solving the “easy problem” for two reasons. First, the “easy problem” can be provisionally solved by tackling bias, harm, and transparency issues. Second, solving the “easy problem” helps solve the “hard problem”, as responsible organizational AI practices create virtuous rather than vicious cycles.