The Unpriced Externality: Toward a Data and Attention ESG (DAESG) Framework for AI Governance

Digital ecosystems collapse the same way physical ecosystems do: through extraction that outpaces regeneration. This paper argues that attention extraction and data-commons depletion constitute unpriced externalities in AI and platform governance, analogous to the unpriced carbon externalities that preceded the emergence of environmental, social, and governance (ESG) accounting. Two extraction vectors are identified: attentional extraction, evidenced by engagement-maximizing design features now subject to regulatory scrutiny, including the European Commission's July 2026 preliminary finding that Meta's Facebook and Instagram breach the Digital Services Act through addictive design; and data-commons extraction, evidenced by model collapse, the documented degradation of AI systems trained on recursively generated synthetic data. The paper argues that both vectors reflect a deeper structural pattern: a decoupling of reward from risk, in which the party capturing the value of extraction is not the party bearing its cost. Building on this diagnosis, the paper proposes DAESG: a three-layer accountability framework, comprising disclosure, standards, and capital allocation, structured to give financial governance functions a basis for treating digital ecosystem extraction as a material, disclosable, and manageable risk. The paper distinguishes DAESG from existing “Digital ESG” (DESG) literature, which addresses how digitalization improves corporate ESG performance rather than the extraction risks of digital ecosystems themselves.