Paper Number

ICIS2026-2588

Paper Type

Completed

Abstract

The safe and responsible development of artificial general intelligence (AGI) is among the most consequential challenges facing organizations and society today. Yet how AGI-pursuing organizations construct and articulate digital responsibility remains not well understood. Drawing on hybrid organization theory, we examine, through a case study of OpenAI and Anthropic, how these organizations manage field-level tensions between development speed and model safety. Organizations respond by deploying a portfolio of five practice types, of which four constitute an organizational responsibility architecture, while the subthemes of ‘market positioning’ appear to serve as legitimacy signal or acceleration-supporting mechanism. The responsibility architecture has a seemingly strong impact on AI safety, weak impact on broader sustainability implications, and is largely internally controlled. These findings reveal structural limits of voluntary self-governance, highlight the need for governance frameworks that extend beyond organizational discretion, and illustrate how engineering overlaps between the commercial and welfare logic can steer responsible behavior.

Comments

06_Societal_impacts

Download

Share

COinS
 
Dec 14th, 12:00 AM

Managing Tensions Between Speed and Safety: Digital Responsibility in Hybrid AGI Organizations

The safe and responsible development of artificial general intelligence (AGI) is among the most consequential challenges facing organizations and society today. Yet how AGI-pursuing organizations construct and articulate digital responsibility remains not well understood. Drawing on hybrid organization theory, we examine, through a case study of OpenAI and Anthropic, how these organizations manage field-level tensions between development speed and model safety. Organizations respond by deploying a portfolio of five practice types, of which four constitute an organizational responsibility architecture, while the subthemes of ‘market positioning’ appear to serve as legitimacy signal or acceleration-supporting mechanism. The responsibility architecture has a seemingly strong impact on AI safety, weak impact on broader sustainability implications, and is largely internally controlled. These findings reveal structural limits of voluntary self-governance, highlight the need for governance frameworks that extend beyond organizational discretion, and illustrate how engineering overlaps between the commercial and welfare logic can steer responsible behavior.

When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.