文本到视频生成模型中的性别偏见：以 Sora 为例
Gender Bias in Text-to-Video Generation Models: A case study of Sora
AI总结：
本文通过分析 Sora 对性别中立及刻板印象提示生成的视频，揭示其将特定性别与行为和职业不当关联的偏见及训练数据根源。
摘要：
文本到视频生成模型的出现彻底改变了内容创作，因为它能够根据文本提示生成高质量视频。然而，对此类模型固有偏见的担忧引发了严格审视，尤其是在性别表征方面。本研究调查了 OpenAI 的 Sora（一款最先进的文本到视频生成模型）中性别偏见的存在。通过分析由多样化的性别中立和刻板印象提示生成的视频，我们发现了显著的偏见证据。结果表明，Sora 不成比例地将特定性别与刻板行为和职业联系起来，这反映出其训练数据中嵌入的社会偏见。
英文摘要：
The advent of text-to-video generation models has revolutionized content creation as it produces high-quality videos from textual prompts. However, concerns regarding inherent biases in such models have prompted scrutiny, particularly regarding gender representation. Our study investigates the presence of gender bias in OpenAI's Sora, a state-of-the-art text-to-video generation model. We uncover significant evidence of bias by analyzing the generated videos from a diverse set of gender-neutral and stereotypical prompts. The results indicate that Sora disproportionately associates specific genders with stereotypical behaviors and professions, which reflects societal prejudices embedded in its training data.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
