What LLMs are doing to conference programs (di.nmfay.com)

In a recent exploration of the impact of large language models (LLMs) on conference programming, insights from PG DATA 2026 revealed significant challenges in evaluating submissions. The ease with which LLMs can generate polished abstracts has led to an uptick in superficial or irrelevant submissions, complicating the traditionally nuanced selection process conducted by program committees. Organizers noted that a not insignificant number of abstracts lacked relevance to the conference focus, indicating that while LLMs can enhance the writing process, they can also dilute the quality of content. This trend raises critical implications for the AI/ML community, particularly concerning the authenticity and reliability of speaker expertise. The reliance on LLMs can mask a speaker's true understanding of their topic, making it difficult for committees to gauge the potential value of presentations. The author suggests that instead of relying on AI to craft or polish submissions, speakers should focus on honing their writing skills, using LLMs primarily as tools for critique rather than content creation. As the landscape evolves, balancing the benefits of AI assistance with the necessity for genuine human expression will be essential for maintaining the integrity and quality of academic discourse.