The Hidden Carbon Cost of Short-Form Video AI: Examining the Sustainability Paradox in Social Media Marketing
Synopsis
This study investigates the sustainability paradox inherent in AI-driven short-form video marketing, where the rapid adoption of generative tools (e.g., Runway, Pika, Sora) increasingly conflicts with corporate Net Zero commitments. Drawing on a systematic literature review (2019-2026) at the intersection of AI ethics, environmental science, and marketing, alongside scenario-based carbon modeling, the analysis demonstrates that video generation consumes approximately 30 times more energy than image creation. Estimates indicate that a mid-sized marketing team generating 2,500 AI videos annually can emit up to 325.5 kg CO₂ - a hidden environmental cost largely obscured by decentralized "shadow AI" practices and the lack of AI-specific sustainability metrics in current KPIs. To address this reporting gap, particularly in light of expanding Scope 3 disclosure requirements under the EU CSRD, the paper introduces the Carbon Per Mille (C-CPM) indicator. By proposing algorithmic greenwashing as a conceptual lens, this research provides an initial academic assessment of the carbon footprint of AI video generation and advocates for the integration of carbon-aware operational strategies.






