Social Sustainability in the Use and Development of Artificial Intelligence

Sammanfattning

The development of artificial intelligence (AI) has transformed how organizations operate, creating new opportunities and responsibilities regarding social sustainability and human rights. As AI tools continue to evolve, questions arise about how social sustainability is interpreted, managed, and governed in practice. This study examines how social sustainability, including human rights, is defined and implemented in AI-related work within a global software organization. The organization operates within a large, established industrial group, and the study aims to analyze how organizational and strategic tensions impact social sustainability in the development of AI. The study is based on semi-structured interviews with 12 respondents from the holding company or one of its subsidiaries; the respondents represent various job roles and 5 countries. The results were then analyzed using stakeholder theory and socio-technical theory. The results clearly show that social sustainability is formally embedded at the group level through international commitments, governance structures, and policies. However, this becomes less visible and increasingly fragmented and informal the closer it gets to the site of AI development and implementation of AI-related processes. At the operational level, social sustainability is managed through compliance and individual responsibility, rather than through strategic governance structures. Responsibility is distributed across several different units and actors without a clear owner, creating an accountability gap. The pace of AI development adds another layer of challenges, as governance structures cannot keep pace with this rapid development. Innovation speed and commercial pressure systematically crowd out social sustainability considerations, and the fear of falling behind consistently outweighs the incentive to pause and govern responsibly. The study contributes to existing research by showing that the development of AI also gives rise to new relationships among stakeholders for organizations that current governance processes do not address. These include employees who experience significant fluctuations in their workload, local communities affected by the server infrastructure, and customers who lack the expertise to review AI results. Furthermore, the accountability gap stems more from governance structures than from intent. The study also provides further support for the socio- technical theory, as evidenced by the finding that the parallel development of technical and social systems does not occur spontaneously, as the organization’s values might suggest, but rather requires clear governance structures across the entire organization. Joint optimization is undermined when the pace of technical development exceeds the ability of social governance structures to evolve at the same pace. The findings suggest that organizations need clearer governance structures, defined ownership, and cross-functional coordination to integrate social sustainability into AI-related work. The employees who use AI also need a broader understanding of how AI systems actually work, what data they are trained on, and how AI tools should be used across all functions.