The United Nations created the 17 Sustainable Development Goals (SDGs) to promote environmental protection, economic growth, and social justice. In this scenario, science is crucial to solving the challenges addressed by the SDGs. SciVal, for example, is a tool that tracks scientific publications related to the SDGs with the support of a team of experts. Aiming to reduce the need for specialized knowledge and to provide a more autonomous tool, this study proposes a multi-label classification model based on natural language processing and recurrent neural networks to map scientific publications to the SDGs. The data used to train and evaluate the model are comprised of manuscript titles acquired from the Scopus database using the SciVal analytics tool, and they are related to 16 of the 17 SDGs. The proposed model is applied to the manuscripts of the Brazilian Congress of Automatics (CBA) 2020 as a way to measure the impact of Automation in the SDG. However, the application of the model is not restricted to a specific scientific field. Instead, it can be applied to any field. In the context of the CBA 2020, results have shown that the papers published in the congress focused on SDGs 7, and 9, which are related to clean energy and industry innovation, respectively. Furthermore, all SDGs were associated with at least one publication, indicating that intelligent automation can contribute in an interdisciplinary way to the SDGs implementation.