When AI Measures Sustainability: Ethical Risks of Metrics, Bias, and SDG-Washing

Artificial intelligence (AI) and business analytics are increasingly promoted as instrumental to achieving the United Nations Sustainable Development Goals (SDGs). Governments, corporations, and international organisations now deploy AI-driven indicators, dashboards, and optimisation systems to monitor sustainability performance, guide policy, and signal social responsibility. Despite this growing reliance, the ethical implications of AI-enabled sustainability analytics remain insufficiently examined. This paper critically investigates whether AI-driven business analytics genuinely advance the SDGs or risk devolving into ethical tokenism through SDG-washing, metric manipulation, and normative bias. Drawing on interdisciplinary literature in AI ethics, sustainability governance, and development studies, the paper argues that many AI-enabled SDG initiatives prioritise measurable indicators over substantive social and environmental outcomes. Key ethical challenges are examined, including misalignment between analytics key performance indicators and SDG intent, difficulties in quantifying social values, and biases embedded within sustainability datasets. The paper proposes an ethical governance framework for SDG-oriented AI analytics that emphasises value alignment, contextual sensitivity, outcome-focused evaluation, and accountability. It concludes that AI can contribute meaningfully to sustainable development only when ethical design and governance mechanisms prevent metric-driven compliance from substituting for genuine progress.