Knowledge, power, creation: How generative AI models disrupt the industry

---

Aron Gyenge

Qbiz, United Kingdom

: J Comput Eng Inf Technol

Abstract

---

Large language and other generative AI models have recently been the subject of much discussion, speculation, and fear about their impact on the future of humanity. These models are still in the early stages of development, and the social consequences of their deployment are difficult to predict, but it is clear that such models pose a serious disruptive force in global markets. Research by Goldman Sachs suggests that these models could automate 300 million jobs, making 7% of the US workforce vulnerable to replacement. Enterprises deploying generative AI models have faced controversies, such as class-action lawsuits by artists against companies like Stability AI and Mid journey, or the ban of ChatGPT in Italy due to data protection concerns. While concerns about the development of generative models are understandable, these cases may set a precedent for banning the use of publicly available data, which could hinder businesses and beneficial uses of data science. However, the core of the problem at hand is not simply a conflict between technocrats and neo-Luddites, but rather an ethical dilemma related to information asymmetry, power dynamics, and exploitation within the framework of information capitalism. Therefore, it is important to consider the social arrangements at the heart of this issue. More specifically, that web-based information societies generate invaluable data for the training of state-of-the-art machine learning models, but the work of these societies does not align with traditional labour theories of capitalist societies. That is, to consider how the disparity between real and perceived labour fuels the commercial and social disruption of generative models.

Biography

---