Roadmap to competitive and socially responsible artificial intelligence

Abstract

The roadmap to competitive and socially responsible artificial intelligence (AI) offers an overview of AI governance drivers and tasks. It is intended for organizations using or planning to use information systems that include AI functionalities, such as machine learning, natural language processing, and computer vision. Responsible AI is still an emerging topic, but legal and stakeholder requirements for AI systems to comply with societally agreed standards are growing. In particular, the European Union’s proposed Artificial Intelligence Act is set to introduce new rules for AI systems used in high-risk application domains. However, beyond binding legislation, soft governance, such as guidelines and ethics principles, already seeks to differentiate between socially responsible and irresponsible AI development and use practices.

The roadmap report begins by laying out its target group, instructions, and structure and then moves on to definitions. Next, we introduce the institutionalization of AI as a necessary background to the consideration of AI governance. The main roadmap section includes a visual representation and explanation of the six key drivers of competitive and socially responsible AI: 1) Movement from AI ethics principles to AI governance 2) Responsible AI commercialization potential and challenges 3) AI standardization 4) Automation of AI governance 5) Responsible AI business ecosystems 6) Stakeholder pressure for responsible AI

The roadmap is followed by a future research agenda highlighting five emerging research areas: 1) operational governance mechanisms for complex AI systems, 2) connections to corporate sustainability, 3) automation of AI governance, 4) future of responsible AI ecosystems, and 5) sociotechnical activities to implement responsible AI. Researchers and research funding bodies play a key role in advancing competitive and socially responsible AI by deepening these knowledge areas.

Advancing socially responsible AI is important because the benefits of AI technologies can be reaped only if organizations and individuals can trust the technologies to operate fairly, transparently, and according to socially defined rules. This roadmap was developed by the Artificial Intelligence Governance and Auditing (AIGA) co-innovation project funded by Business Finland during the years 2020 to 2022. The roadmap was cocreated by researchers, company practitioners, and other AIGA project stakeholders.