Why conferences matter when AI reads and writes papers: a UIST reflection
plopesresearch · x · 2026-11-07
Prof. Masahiko Inami reflects after UIST on what conferences are for in an era when AI reads and writes papers. Many submissions now contain probable AI-hallucinated citations ("HalluCitations"), reviewers feel like unpaid LLM verifiers, and "a conference of AI, by AI, for AI" is becoming plausible.
He argues for keeping human-scale conferences: communities small enough that faces and names connect, where unfinished ideas can be brought; the Kōhaku-like value of single-track sessions, where forced exposure to distant fields can change your own research questions and creates shared context for cross-disciplinary conversation. UIST grew 33% to 1259 submissions across 5 parallel tracks, making shared listening time scarcer; mixers like Korea Night and Japan Lunch remain key for senior researchers introducing students. His conclusion: if conferences only efficiently produce, review and distribute papers, humans lose reasons to gather — their lasting value is serendipitous encounters and co-growing questions.
More from AGI Musings
- Once AI can simulate a human cell, aging becomes a software problem — rand_longevity · 2026-11-07
- Data centers framed as 'killing millions', AI fearmongering sparks pushback — Dr_Singularity · 2026-11-07
- Gary Marcus: many data centers will fail once local AI models take over inference — GaryMarcus · 2026-11-07
- When AI customer service judges your appeal with its own fraud flag, who's accountable? — Creamy-And-Crowded · 2026-11-07
- Is generative drift a fundamental flaw in AI? Reddit debates agent reliability — Peacefulhuman1009 · 2026-11-07
- Creators argue AI writing is flattening prose into a monoculture, betting on human connection — _akpiper · 2026-11-07