Content Moderation in TV Search: Balancing Policy Compliance, Relevance, and User Experience

TL;DR

This paper proposes an LLM-based monitoring layer for TV search content moderation, aiming to improve safety, relevance, and user trust by detecting unintended content and refining search results.

Contribution

It introduces a novel LLM-driven moderation layer that enhances content filtering and feedback integration in TV search systems.

Findings

Improved detection of unintended content

Enhanced user trust and satisfaction

Refined search retrieval through feedback

Abstract

Millions of people rely on search functionality to find and explore content on entertainment platforms. Modern search systems use a combination of candidate generation and ranking approaches, with advanced methods leveraging deep learning and LLM-based techniques to retrieve, generate, and categorize search results. Despite these advancements, search algorithms can still surface inappropriate or irrelevant content due to factors like model unpredictability, metadata errors, or overlooked design flaws. Such issues can misalign with product goals and user expectations, potentially harming user trust and business outcomes. In this work, we introduce an additional monitoring layer using Large Language Models (LLMs) to enhance content moderation. This additional layer flags content if the user did not intend to search for it. This approach serves as a baseline for product quality assurance, with collected feedback used to refine the initial retrieval mechanisms of the search model, ensuring a safer and more reliable user experience.