Validating a Gold Standard for Measuring Political Toxicity and Incivility in the Digital Sphere

About the project

**Principal Investigador:** Bastián González-Bustamante **Research Team:** Sebastian Rivera This project aims to create and validate a gold standard for a set of machine learning algorithms and language models that measure political toxicity and incivility in the digital sphere. For this purpose, we conducted manual data labelling to validate and evaluate deep learning algorithms and Large Language Models (LLMs) applied to protest events in Latin America and to the digital interactions that occurred during the functioning of the Constitutional Convention in Chile. The creation of this reference standard will constitute an empirical contribution that will make it possible to assess the relevance and coherence of the results obtained by applying generative transformers and other machine learning techniques. This constitutes a relevant contribution in terms of algorithmic transparency and artificial intelligence in the social sciences.