Numerous social movements (SMs) around the world help support the UN's Sustainable Development Goals (SDGs). Understanding how key events shape SMs is key to the achievement of the SDGs. We have developed SMART (Social Media Analysis & Reasoning Tool) to track social movements related to the SDGs. SMART was designed by a multidisciplinary team of AI researchers, journalists, communications scholars and legal experts. This paper describes SMART's transformer-based multivariate time series Discourse Evolution Engine for Predictions about Social Movements (DEEP) to predict the volume of future articles/posts and the emotions expressed. DEEP outputs probabilistic forecasts with uncertainty estimates, providing critical support for editorial planning and strategic decision-making. We evaluate DEEP with a case study of the #MeToo movement by creating a novel longitudinal dataset (433K Reddit posts and 121K news articles) from September 2024 to June 2025, which is publicly available for research purposes.

Large Language Models (LLMs) have demonstrated capabilities across a wide range of tasks, including natural language understanding, language generation and generating basic reasoning. To improve the accuracy and reliability of these models, various prompting, augmentation, and fine-tuning strategies have been employed. In this paper, we use context augmentation and Chain-of-Thought (CoT) instructions to generate legal arguments for specific landlord-tenant problems using OpenAI’s GPT-4o. We test this approach using ten hypothetical landlord-tenant scenarios, five provided by a legal aid organization, four generated with Anthropic’s Claude, and one crafted by us. We evaluate each argument for accuracy, factuality, comprehensiveness, and relevance to the scenario. This method of generating legal reasoning with LLMs offers the advantage of being verifiable by legal professionals, while also providing valuable assistance to laypersons in drafting documents such as demand letters, which can help expand access to justice.