Preprint
Large Language Models

When large language models meet personalization: perspectives of challenges and opportunities

Jing Chen(Hong Kong University of Science and Technology), Zheng Liu(Beijing Academy of Artificial Intelligence), Xu Huang(University of Science and Technology of China), Chenwang Wu(University of Science and Technology of China), Qi Liu(University of Science and Technology of China), Gangwei Jiang(University of Science and Technology of China), Yuanhao Pu(University of Science and Technology of China), Yuxuan Lei(University of Science and Technology of China), Xiaolong Chen(University of Science and Technology of China), Xingmei Wang(University of Science and Technology of China), Kai Zheng(University of Electronic Science and Technology of China), Defu Lian(University of Science and Technology of China), Enhong Chen(University of Science and Technology of China)
June 28, 2024World Wide Web348 citations
arXivPDFDOI

348

Citations

13

Influential Citations

World Wide Web

Venue

2024

Year

Abstract

Abstract The advent of large language models marks a revolutionary breakthrough in artificial intelligence. With the unprecedented scale of training and model parameters, the capability of large language models has been dramatically improved, leading to human-like performances in understanding, language synthesizing, common-sense reasoning, etc. Such a major leap forward in general AI capacity will fundamentally change the pattern of how personalization is conducted. For one thing, it will reform the way of interaction between humans and personalization systems. Instead of being a passive medium of information filtering, like conventional recommender systems and search engines, large language models present the foundation for active user engagement. On top of such a new foundation, users’ requests can be proactively explored, and users’ required information can be delivered in a natural, interactable, and explainable way. For another thing, it will also considerably expand the scope of personalization, making it grow from the sole function of collecting personalized information to the compound function of providing personalized services. By leveraging large language models as a general-purpose interface, the personalization systems may compile user’s requests into plans, calls the functions of external tools (e.g., search engines, calculators, service APIs, etc.) to execute the plans, and integrate the tools’ outputs to complete the end-to-end personalization tasks. Today, large language models are still being rapidly developed, whereas the application in personalization is largely unexplored. Therefore, we consider it to be right the time to review the challenges in personalization and the opportunities to address them with large language models. In particular, we dedicate this perspective paper to the discussion of the following aspects: the development and challenges for the existing personalization system, the newly emerged capabilities of large language models, and the potential ways of making use of large language models for personalization.

Analysis

Why This Paper Matters

This perspective paper arrives at a critical juncture where large language models (LLMs) are demonstrating unprecedented capabilities in understanding, reasoning, and language synthesis. The authors argue that these advances will fundamentally reshape personalization systems, which have traditionally been passive filters (e.g., recommender systems, search engines). By positioning LLMs as an active interface, the paper opens up new possibilities for user engagement that is natural, interactive, and explainable.

The significance lies in its forward-looking synthesis: it connects the rapid evolution of LLMs with the well-established challenges of personalization, such as cold-start, sparsity, and user trust. The paper does not merely list opportunities but provides a structured vision of how LLMs can serve as a general-purpose orchestrator—compiling user intents, invoking external tools, and delivering end-to-end personalized services. This perspective is valuable for both researchers and practitioners seeking to align LLM capabilities with real-world personalization needs.

Technical Contributions

  • Paradigm Shift: Moves from passive information filtering to active, proactive user engagement, where LLMs can explore user requests and deliver information in a conversational manner.
  • Expanded Scope: Broadens personalization from collecting personalized information to providing compound personalized services by integrating external tools (e.g., search engines, calculators, APIs).
  • LLM as Interface: Proposes using LLMs as a general-purpose interface that compiles user requests into executable plans, calls external functions, and integrates outputs to complete tasks.
  • Challenges and Opportunities: Systematically reviews existing personalization challenges (e.g., data sparsity, cold-start, explainability) and maps them to LLM capabilities (e.g., few-shot learning, reasoning, natural language generation).

Results

As a perspective paper, no experimental results or quantitative metrics are provided. The contribution is conceptual: it identifies key research directions and potential architectures for LLM-driven personalization. The paper's impact is measured by its citation count (348) and its role in framing future work.

Significance

The broader impact of this paper is its role in catalyzing a new research direction at the intersection of LLMs and personalization. By articulating a clear vision of how LLMs can transform user-system interaction, it provides a roadmap for building more engaging, transparent, and capable personalization systems. This could influence the design of next-generation recommender systems, conversational AI, and intelligent assistants, ultimately leading to more human-centric AI experiences.

large-language-modelspersonalizationrecommender-systemshuman-ai-interactioninformation-filteringuser-engagementexplainable-ai
Large Language ModelsReasoning

Related Papers