The field of natural language recommendation is moving towards more transparent and user-controlled approaches, with a focus on leveraging large-scale datasets and advanced machine learning techniques to improve recommendation accuracy. Researchers are exploring new methods for embedding civic values into news recommendation systems, such as using large-scale audience evaluations to generate value-based labels. Additionally, there is a growing interest in developing more nuanced and fine-grained opinion mining approaches, including the use of fuzzy logic algorithmic approaches to classify opinions into different granularity levels. Noteworthy papers in this area include: SciNUP, which introduces a novel synthetic dataset for scholarly recommendation that leverages authors' publication histories to generate NL profiles and corresponding ground truth items. Multimodal Item Scoring for Natural Language Recommendation via Gaussian Process Regression with LLM Relevance Judgments, which proposes a new approach to natural language recommendation that uses Gaussian Process Regression with LLM relevance judgments to better capture the potential multimodal distribution of the relevance scoring function. ORBIT, which introduces a unified benchmark for consistent and realistic evaluation of recommendation models, featuring a standardized evaluation framework and a new webpage recommendation task.