Abstract for: From hybris to eudaimonia?: The problem of the AI innovation pathway and its realignment in an unequal and finite world.

Innovation is a key enabler of societal development. But do our current ways help move our society in desirable directions? The emerging GenAI/LLM industry begs this question, as it promises drastic innovation across sectors, but is also growing, attracting and consuming resources, and altering human attention and practices at unprecedented rates, with vast societal consequences. Recognizing AI as a cultural and social technology with wide reach this paper performs an integrative archival-based analysis to examine the main mechanisms underlying the present AI innovation pathway and its various social-environmental impacts—within realms ranging from resource extraction and consumption to greenhouse gas emissions, to labor exploitation and displacement, inequalities, education systems, mis- and disinformation, to global security. The broad boundary analysis characterizes not only diverse areas of impact and, where possible, their extents, but, using causal loop diagramming, also the underlying dynamics driving these, including self-perpetuating dependence on growth, exuberant expectations, and crowding out of other promising developments. Together this demonstrates an industry at risk of locking into a pathway with near- and longer-term environmental irreversibilities and erosion of valued collective practices, and institutions. The paper discusses courses of action for realigning the AI innovation path, to one with drastically improved positive and decreased negative consequences. Besides efforts involving collective governance, promotion and protection of responsible AI practices, and education, the findings suggest a “precautionary principle of speed” across levels, needed to help achieve mindset changes about and societal actions on AI in a direction that is consistent with societal thriving in a resource-constrained world.