The origins of generative AI (GAI) can be traced back to the 1950s, when Alan Turing proposed a test to determine whether a machine could be perceived as intelligent enough to generate responses to questions in a way indistinguishable from a human. Later, in the 1970s, researchers developed more advanced models capable of producing more realistic and coherent outcomes. Contemporary GAI models are based on state-of-the-art neural network architectures. They combine such architectures to develop large models that outperform existing benchmarked ones. Contemporary GAI solutions can produce large amounts of contextual outputs on any specific topic. They are highly trained and sophisticated, enabling users to produce various types of AI-generated content. Although GAI has been around for a while, recent developments have brought the potential of such solutions to the forefront. In particular, LLMs have the potential to transform the way we develop textual content and communicate with one another.
The ongoing discourse on GAI seems to extol the promises of AI and the dangers. Our goal for this Special Issue is to offer a careful examination of the challenges faced in managing this powerful set of technologies for individuals, organizations, and society. Many of the challenges around GAI concern data. As per a Forbes report, over 90% of internet data will be produced by GAI models, triggering serious concerns about harmful and abusive content generation. Most current GAI-triggered use involves chat-based digital assistants. While the outcome of GAI in these digital assistant-based applications is indeed remarkable, their effectiveness depends on the level of task specificity and the need for information synthesis.
At the individual level, a number of challenges exist on how to effectively use GAI to augment individual productivity. For instance, how can GAI-based interactions positively or negatively affect customer experiences, how can GAI augment (vs. replace) human skills, and broader questions of how over-reliance on GAI systems may adversely impact the cognitive inability of users and learners. At the organizational level, there are many challenges around governance. For instance, how can we govern the quality of content by GAI, how can the adoption of GAI lead to disruption, how do we set up appropriate governance structures to manage GAI projects, and how can we avoid unintended consequences of GAI adoption in firms? At the societal level, there are extensive challenges around misinformation, bias, and privacy. Our broad goal for the special issue is to attract papers that articulate the challenges theoretically and study them empirically, while making a strong contribution to the theory and practice in the deployment of GAI.