Abstract Should we adopt broad, anticipatory regulatory frameworks for AI to address its wide-ranging societal impacts? This Article argues that such a universal approach is misguided. Instead, it contends that we should adopt a two-track regulatory strategy that maintains agnosticism about AI’s uncertain long-term trajectory while implementing targeted interventions to address specific, documented harms in the present. The Article develops this argument through a case study of generative AI systems in digital media, examining how these technologies are reshaping the algorithmic dynamics of the public sphere. Tracing the ways in which AI-powered content generation and curation are amplifying long-standing challenges like the erosion of shared epistemic foundations, the entrenchment of echo chambers, and the concentration of informational power, the Article highlights the limitations of the EU’s horizontal AI Act and the potential of China’s more focused, domain-specific approach to regulating these harms. Drawing on this analysis, it makes the case for a hybrid governance model that combines targeted oversight of AI’s known dangers with a continuing spirit of regulatory humility and openness toward its transformative potential. In an age of rapid technological change, this balancing act offers a path beyond the false choice between reactive overregulation and reckless laissez-faire - one that remains attuned to AI’s contextual nuances while advancing the public interest.
Gilad Abiri (Wed,) studied this question.