The promise of “sustainable AI” (“green AI”) is increasingly espoused by researchers, governments, practitioners, and Big Tech alike. While often well-intentioned, this discourse faces a critical danger: its co-option by powerful industry actors. Discussions surrounding AI’s environmental impact are being deliberately narrowed to resource consumption and ethical considerations—framed as technical problems solvable through innovation.
This approach mirrors risk-benefit dichotomies by effectively ignoring deeply embedded systemic injustices perpetuated by the current AI ecosystem and its reinforcement of existing power structures. Furthermore, framing AI sustainability as achievable through technological progress ironically justifies the deployment of AI-driven technologies in resource management, positioning Big Tech as a provider of data-focused solutions. Even though these data-intensive solutions often promise sustainable extraction, they accelerate environmental degradation and exacerbate global distribution injustices.
This talk will demonstrate how this narrow focus diverts attention from the root causes of unsustainability relating to AI, challenging the assumption that technological innovation alone can resolve complex socio-ecological issues and advocating for a more holistic, justice-oriented approach to AI sustainability.