谷歌旗下的新型AI系统Gemini近期推出的人物图像生成功能被爆出存在算法歧视问题。部分用户反馈称,Gemini生成的人物图像过于刻意追求种族和性别的多元化,与实际史实不符,存在反向歧视。作为回应,谷歌宣布将暂停这一功能,并表示将尽快推出改进版本。这一事件再次引发了人工智能可能带来的算法偏见和歧视问题的讨论。一方面,算法训练过程中难免会反映数据本身所存在的偏见;另一方面,盲目追求政治正确,也有可能导致另一种形式的歧视。如何在二者间找到平衡,仍是一个有待探索的难题。
Title: Google AI Character Image Generation Faces Controversy Over Reverse Discrimination
Keywords: AI discrimination, pursuit of diversity, function suspension
News content: Gemini, Google’s new AI system, has recently sparked controversy over algorithmic discrimination in its character image generation feature. Some users complained that the generated images were overly pursuing racial and gender diversity in an unrealistic way, amounting to reverse discrimination. In response, Google announced the suspension of this feature and promised to release an improved version soon. This incident has reignited discussions around the biases and discrimination AI algorithms may introduce. On one hand, biases existing in the training data can inevitably creep into models; on the other hand, blind pursuit of political correctness can also lead to another form of discrimination. Striking a balance between the two remains an open challenge to explore.
【来源】https://www.ithome.com/0/751/457.htm
Views: 1