Jonathan Teubner (Harvard U Institute Quantitative Social Science) and Ronald Ivey (Harvard U Institute Quantitative Social Science) have posted “Social AI and Human Connections: Benefits, Risks and Social Impact” on SSRN. Here is the abstract:
Since the mid-2000s, the widespread adoption of social media has coincided with a significant decline in face-to-face socialization, raising concerns about its broader impact on mental health, community well-being, and democratic governance. This paper argues that the failure of institutions to effectively respond to the social consequences of digital technology provides an important lesson as we confront a new wave of transformation brought on by artificial intelligence.
To understand this challenge, we identify and analyze two intersecting trends: (1) the continuing erosion of in-person social connectedness in the United States, and (2) the rapid advancement and adoption of conversational AI systems—particularly those designed to simulate social and emotional engagement. These systems, which we refer to as “Social AI” (as defined by Shevlin, 2024) are increasingly positioned to fulfill roles traditionally occupied by human relationships.
Drawing on a review of recent literature, expert interviews, a Salon with leading technologists and scholars of human flourishing, and webinars with Social AI researchers, explore the question: How might we design AI systems for social connectedness and human flourishing?
The paper presents a systemic framework for understanding Social AI and its effects on human social capabilities, and it outlines stakeholder-driven recommendations across five domains: (1) AI systems, (2) AI businesses, (3) AI markets, (4) political systems, and (5) cultural and social systems. We propose this as a foundation for future research, impact evaluation, and policy analysis aimed at ensuring AI supports, rather than undermines, human flourishing.