Socially responsible digital innovation and its implications for large language models in AI development

This paper explores the connection between socially responsible innovation and emerging artificial intelligence-based digital technologies using large language models (LLMs) as an example. While socially responsible licensing (SRL) has been examined in sectors such as health and medicine, its application to digital technologies and AI remains underexplored. To address this gap, we review the guiding principles of SRL to promote sustainable development of artificial intelligence (AI) tools for long-term societal benefit and impact. We conducted a review of the literature on SRL, synthesizing findings from 42 articles to identify its guiding principles, application areas, and its relevance in the context of LLMs. Our analysis reveals that the principles of SRL are highly relevant to information systems development and to the tech industry, which is increasingly concerned not only with commercialization but also with the social impact of its innovations. The results further highlight that health and medicine are the primary sectors where SRL has been applied to date. The paper identifies key concepts of SRL (e.g. university licensing, global access clauses, non-exclusive licensing) to address key themes for future exploration in AI development: non-exclusivity for public benefit, data access and equity, and transparency and accountability. We conclude with a call to action for policymakers to enforce SRL principles in AI development to ensure these technologies serve the public good.