PulseAugur
EN
LIVE 12:01:16

AI models tackle humor generation and understanding across text and video

Researchers are developing new methods to evaluate and generate humor using AI, addressing the subjective nature of what makes something funny. One approach involves a "generate-many, select-best" strategy using preference modeling, which ranked highly in SemEval-2026 humor generation tasks. Another study analyzed dark humor and regular humor on YouTube Shorts, finding that dark humor elicits more mixed and sometimes toxic audience reactions. Additionally, a new benchmark called v-HUB has been created to assess multimodal LLMs' ability to understand humor in videos, highlighting that incorporating audio improves comprehension. AI

IMPACT New benchmarks and methodologies are emerging to evaluate AI's nuanced understanding and generation of humor, pushing the boundaries of multimodal and constrained AI capabilities.

RANK_REASON Multiple research papers introduce new benchmarks, datasets, and methodologies for evaluating and generating humor using AI.

Read on arXiv cs.AI →

AI-generated summary · Google Gemini · from 4 sources. How we write summaries →

COVERAGE [4]

  1. arXiv cs.AI TIER_1 English(EN) · Alexey Tikhonov, Alexey Ivanov ·

    lmfaoooo at SemEval-2026 Task 1: Humor Is an Audience. Preference Modeling for Constrained Humor Generation

    arXiv:2606.00022v1 Announce Type: cross Abstract: Humor generation remains difficult not only because producing fluent, novel jokes is hard, but because "funny" is audience-dependent and supervision is noisy -- preferences vary with audience, context, and culture, and annotator a…

  2. arXiv cs.AI TIER_1 English(EN) · Sydney Johns, Sanjeev Parthasarathy, Shantnu Bhalla, Vaibhav Garg ·

    When Jokes Cross the Line: Analyzing Regular Humor and Dark Humor in YouTube Shorts

    arXiv:2606.00046v1 Announce Type: cross Abstract: Video platforms such as YouTube have reshaped how users engage with entertainment and information, emphasizing brief, highly engaging content such as Shorts. Within this ecosystem, certain content occupies a gray area where it rem…

  3. arXiv cs.AI TIER_1 English(EN) · Zhengpeng Shi, Yanpeng Zhao, Jianqun Zhou, Yuxuan Wang, Qinrong Cui, Wei Bi, Songchun Zhu, Bo Zhao, Zilong Zheng ·

    v-HUB: A Benchmark for Video Humor Understanding from Vision and Sound

    arXiv:2509.25773v3 Announce Type: replace-cross Abstract: AI models capable of comprehending humor hold real-world promise -- for example, enhancing engagement in human-machine interactions. To gauge and diagnose the capacity of multimodal large language models (MLLMs) for humor …

  4. arXiv cs.CL TIER_1 English(EN) · Edward Ajayi, Prasenjit Mitra ·

    HumorRank: A Tournament-Based Leaderboard for Evaluating Humor Generation in Large Language Models

    arXiv:2604.19786v2 Announce Type: replace Abstract: Humor remains difficult to evaluate in large language models (LLMs) because what makes a response funny is subjective, comparative, and shaped by interacting comedic mechanisms rather than a single scalar property. Existing humo…