据 BBC 和 The Verge 报道，谷歌正在紧急处理其 Gemini 大模型中的一个意外问题，该问题导致其文生图功能过度强调“多元化”，甚至在某些情况下生成与史实不符的图像。用户反馈指出，Gemini 在创建图像时倾向于避免白人形象，而倾向于展示各种性别和种族的人物，这引发了对模型偏见的担忧。
谷歌已经宣布，作为纠正措施，将暂时停止 Gemini 的人物图像生成服务，以集中力量解决这一与多元化呈现相关的问题。该公司表示，理解并重视公正性和准确性在人工智能中的重要性，承诺将尽快推出一个修复后的改进版本。
谷歌的这一决定凸显了科技巨头在推进人工智能技术时，对避免算法偏见和确保内容真实性的重视。随着人工智能在媒体和信息生成中的角色日益增强，如何在创新与责任之间找到平衡，成为业界亟待解决的挑战。谷歌的行动预示着未来可能会有更多类似的调整，以确保技术的发展既多元化又准确无误。
英语如下：
News Title: “Google Pauses Gemini Image Generation: Pursuing Diversity at the Cost of Accuracy, Set to Resume After Fix”
Keywords: Google fix, Gemini controversy, diverse imagery
News Content:
In a move to address unintended consequences, Google has temporarily halted the character image generation aspect of its Gemini model, as reported by BBC and The Verge. The issue stems from the model’s excessive emphasis on “diversity,” which in some instances led to the creation of historically inaccurate images. Users have noted that Gemini tended to avoid depicting white individuals while favoring characters from various genders and ethnicities, raising concerns about potential biases in the model.
Google has announced it will suspend the service to concentrate on rectifying this issue related to the portrayal of diversity. The company acknowledges the significance of fairness and accuracy in AI and vows to roll out an improved version as soon as possible.
This decision underscores the tech giant’s commitment to mitigating algorithmic biases and ensuring content authenticity in the development of AI technologies. As AI’s role in media and information generation grows, striking a balance between innovation and responsibility poses a pressing challenge for the industry. Google’s action signals that more such adjustments may follow, aiming to guarantee that technological advancements promote diversity without compromising factual correctness.
【来源】https://www.ithome.com/0/751/457.htm
Views: 1