Konuşmacılar
Açıklama
This paper examines the potential threats of AI to human dignity in social science research through the lens of Scanlonian contractualism, offering both an explanation of the moral risks and a framework for addressing them. Human dignity is defined as the inherent worth of individuals, grounded in their capacity for rational moral agency. AI systems in social science research pose unique challenges to this concept by potentially undermining human agency, reducing individuals to mere data points, or reinforcing societal hierarchies.
Drawing on T.M. Scanlon's contractualism, particularly its ex ante interpretation, the article argues that AI applications in social science threaten human dignity when they can be reasonably rejected by those affected. The analysis considers three types of AI-related risks in research contexts: those affecting morally salient groups, less common groups, and random individuals.
The paper proposes that for AI use in social science to be morally permissible, it must impose only "acceptable risks on human dignity." This requires that research participants and subjects expect to benefit from the AI system's application. Strategies are offered for researchers to mitigate risks: (1) reducing expected harms and (2) increasing expected benefits.
The article presents case studies from social science research to illustrate these concepts and their application. It concludes by offering guidelines for social scientists, research institutions, and policymakers to ensure that AI systems in research respect human dignity while advancing scientific knowledge.
This interdisciplinary approach provides a nuanced framework for AI ethics in social science, emphasizing human dignity beyond traditional fairness considerations. It aims to inform research practices and policies, ensuring responsible AI use in social science while acknowledging the need for technological progress in the field.
Keywords: human dignity, AI, contractualism, T.M. Scanlon, social risk
Institution / Affiliation / Kurum
Kadir Has Üniversitesi
| Presentation language / Sunum Dili | EN (English) |
|---|---|
| Disciplines / Disiplinler | Ethics / Etik |
| E-mail / E-posta | huseyink@gmail.com |
| ORCID ID | 0000-0002-9267-4891 |
| Country / Ülke | Türkiye |
