Working Papers

This paper proposes the use of synthetic training data generated by large language models
to improve machine learning SDG classifiers. It shows that supplementing existing training data with
synthetic data produced by the ChatGPT tool improves the performance of the SDGClassy classifier.
This addition of synthetic data is especially useful in building SDG classifiers given the limited availability
of properly labeled data and the complex, interconnected nature of the SDGs. Synthetic data thus enables
more effective machine-learning applications in this context.

Research suggests that achieving the Sustainable Development Goals (SDGs) can open up market opportunities worth USD 12 trillion in the four largest sectors that represent 60 per cent of real economy – food and agriculture; cities; energy and materials; and health and well-being. While the concept of the SDGs creating a win-win situation for all is growing increasingly trendy, further evidence needs to be accumulated to better chart the important discourse on the private sector’s engagement with the SDGs. To this end, this paper aims to shed light on three questions: (i) How is the private sector currently engaging with the SGDs in these sectors?; (ii) What are the key areas of…