Generative AI in practice: An umbrella review of risks, benefits, ethics, and future directions across major domains

Generative artificial intelligence has rapidly evolved, influencing diverse fields such as healthcare, education, and technology. While the benefits of GAI, such as automation, decision-making support, and personalization, are widely recognized, concerns regarding bias, misinformation, ethical implications, and regulatory gaps persist. We conducted an umbrella review to synthesize existing literature on GAI, focusing on key research topics, methodological trends, associated risks and benefits, ethical considerations, AI literacy, and regulatory strategies. The analysis of N = 170 studies reveals that the majority of research centers on health, technology, and education. Importantly, the literature presents a balanced discourse of risks and benefits. However, the framing of risks and benefits varied by topical domain. Furthermore, ethical concerns remain comparatively underexplored and AI literacy only received little attention, especially in the health and technology domains. Also, there is an overall need for more empirical evaluations of AI adoption, interdisciplinary perspectives, and enhanced ethical guidelines. Implications for the interdisciplinary study of GAI are discussed. • An umbrella review on generative AI finds that health, technology, and education are the main foci. • There is a balanced discourse of risks and benefits. • However, the framing of risks and benefits varies by topical domain. • Ethical concerns remain comparatively underexplored. • The importance of AI literacy is not fully recognized in the literature.