Send message to Author

Short abstract:

This paper studies how 'data professionals' — data scientists and actuaries within a disability insurer — legitimize their work amidst public controversies of algorithms. It contributes to the classification literature by showing the critical role of professional morality in shaping market dynamics.

Long abstract:

AI has revolutionized the way markets classify and sort individuals, offering substantial economic gains while posing significant societal risks due to unexpected impacts on individuals' life chances. This paper studies how 'data professionals' — data scientists and actuaries within a disability insurer — legitimize their work amidst these challenges. Using an ethnographic observations and interviews, the study explores the moral judgments each professional groups makes in their distinct risk assessment system. Actuaries focus on calculating premiums through traditional risk assessments, while data scientists leverage machine learning to predict claims risk, highlighting a crucial shift toward technological solutions in claims prevention. The findings reveal a stark moral divergence between these groups. Data scientists are outspoken on 'data ethics’ and have designed internal ethical guidelines that stress the necessity for developing fair and accurate models aimed at helping customers in need. In contrast, actuaries remain rather silent on their ethical considerations but adhere to notions of 'actuarial fairness', favoring explainable, simplistic models that group individuals into risk categories, thereby supporting solidarity through pricing. By comparing these professional morals, the paper shows how different moral viewpoints inform the outcomes of market classifications – while data scientists produce personalized risk scores, actuarial work produces group-based classifications. This paper contributes to the classification literature in economic sociology by signaling the critical role of professional morality in shaping market dynamics and societal impacts. It points to the continuing relevance of the professional background of ‘data professionals’ instead of focusing exclusively on the morals of algorithms.