Track: Type A (Regular Papers)
Keywords: Foundation Models, Applied AI, AI in media
Abstract: The rapid advancement of Generative AI (GenAI) has made AI a central focus for many organizations. They are keen to explore how GenAI can optimize their operations. Yet, they recognize that introducing GenAI influences their employ-ees’ jobs and customers’ experience. Hence, implementing GenAI in organiza-tions introduces a variety of new technical, practical, and ethical challenges, re-quiring careful decision-making. This study examines how three media organiza-tions approach these challenges when developing and implementing applications based on Foundation Models: large generic GenAI-models that serve as a foun-dation for task-tailored models and applications. Through qualitative analysis of data collected in interviews and focus group sessions, organizations’ considera-tions in decision-making concerning Foundation-Model based applications were identified. Five central themes were determined in this decision-making: use case selection, handling data, choices in models and infrastructure, evaluation, and in-tegration and future-readiness in the organization. Adopting GenAI within organ-ization-specific applications proved to be a multidimensional balancing act. Three dimensions stood out as especially challenging: 1) weighing short-term return on investment against future requirements for flexible and inclusive applications; 2) weighing ease of use and maintenance that large technology providers offer against transparency and control offered by open-source models; and 3) consider-ing impact on human work, by addressing not only potential time savings but al-so possible job obsolescence and the creation of new tasks. The study under-scores the importance of developing clear frameworks and best practices that help assess the diverse implications of adopting GenAI-based applications in real-world organizational settings.
Serve As Reviewer: ~Sietske_Tacoma1
Submission Number: 45
Loading