AI as Entertainment

TL;DR

The paper addresses the mismatch between AI evaluation focused on intelligence and the rising prominence of AI in entertainment. It advocates a shift toward thick entertainment, drawing on humanities concepts to assess cultural meaning, friction, and ambiguity in AI-generated content. By synthesizing industry data, surveys, and theory, it argues that entertainment could become a dominant revenue model for AI firms and that current metrics fail to capture beneficial cultural outcomes. A constructive framework for evaluating AI as entertainment is proposed, with implications for policy, governance, and design to shape AI’s societal impact. The work underscores the urgency of proactively studying and guiding AI-driven cultural processes to avoid detrimental, profit-driven trajectories.

Abstract

Generative AI systems are predominantly designed, evaluated, and marketed as intelligent systems which will benefit society by augmenting or automating human cognitive labor, promising to increase personal, corporate, and macroeconomic productivity. But this mainstream narrative about what AI is and what it can do is in tension with another emerging use case: entertainment. We argue that the field of AI is unprepared to measure or respond to how the proliferation of entertaining AI-generated content will impact society. Emerging data suggest AI is already widely adopted for entertainment purposes -- especially by young people -- and represents a large potential source of revenue. We contend that entertainment will become a primary business model for major AI corporations seeking returns on massive infrastructure investments; this will exert a powerful influence on the technology these companies produce in the coming years. Examining current evaluation practices, we identify a critical asymmetry: while AI assessments rigorously measure both benefits and harms of intelligence, they focus almost exclusively on cultural harms. We lack frameworks for articulating how cultural outputs might be actively beneficial. Drawing on insights from the humanities, we propose "thick entertainment" as a framework for evaluating AI-generated cultural content -- one that considers entertainment's role in meaning-making, identity formation, and social connection rather than simply minimizing harm. While AI is often touted for its potential to revolutionize productivity, in the long run we may find that AI turns out to be as much about "intelligence" as social media is about social connection.

Figures (1)

• Figure 1: Contemporary evaluation protocols tend to focus on the benefits of intelligence and the harms of culture. Here we show a rough conceptual sketch organizing a subset of evaluations from a survey by Burden and colleagues (2025) burden2025paradigms along two axes: harm vs benefit and cultural vs intelligent performance. We distinguish between intelligent performance as the ability to execute some sort of goal-oriented, instrumental process and cultural performance as the ability to abide by contextually relevant norms or accord with subjective human judgments. For example, multimodal test suites patraucean2023perception, cogsci tasks coda2024cogbench, Humanity's Last Exam phan2025humanity, and benchmarks for automated AI researchers si2024can evaluate the potential benefits of intelligence. By contrast, safety evaluations span both culture and intelligence but by definition focus exclusively on harms. This figure is not intended to serve as an authoritative delineation between culture and intelligence or to provide a comprehensive account of extant evaluation protocols. Rather, it meant to illustrate our claim that there is a relative paucity of AI evaluations that look at the benefits of culture.