谷歌计划数周内重新上线 Gemini 人像生成功能
据路透社当地时间 26 日报道，谷歌 DeepMind 首席执行官德米斯·哈萨比斯在 MWC 2024 的小组讨论会上表示，谷歌已下线 Gemini 的人像生成功能，目前正在修复问题，并希望在未来几周内重新上线。
Gemini 是谷歌开发的一种人工智能模型，可以生成逼真的人像。去年 12 月，谷歌首次发布了 Gemini，但随后不久便将其下线，原因是该模型生成了带有性别和种族偏见的图像。
哈萨比斯表示，谷歌已采取措施解决 Gemini 的偏见问题，并对该模型进行了重新训练。他表示，新版本的人像生成功能将更加准确和包容。
谷歌发言人证实了哈萨比斯的言论，并表示公司致力于开发负责任的人工智能技术。发言人表示：“我们正在努力确保 Gemini 符合我们的道德准则，并且不会被用于有害或歧视性的目的。”
Gemini 的重新上线将标志着谷歌在生成式人工智能领域取得的重大进展。生成式人工智能是一种能够创建新内容的人工智能，例如图像、文本和音乐。
专家认为，Gemini 的重新上线将为创意专业人士和内容创作者提供新的工具。然而，他们也警告说，重要的是要负责任地使用生成式人工智能，并解决其潜在的偏见和滥用问题。
英语如下：
**Headline:** Google’s Image-Generating Tool to Return in Weeks After Fixes
**Keywords:** Google model, image generation, fixes
**Body:**
Google plans to relaunch its Gemini image-generating tool within weeks, the companysaid Sunday.
Google DeepMind CEO Demis Hassabis said at a panel discussion at MWC 2024 that Google had taken down Gemini’s image generation feature and was working to fix the issue, with plans to relaunch it in the coming weeks, Reuters reported.
Gemini is an artificialintelligence model developed by Google that can generate realistic images of people. Google first released Gemini in December, but took it down shortly after because the model was generating images with gender and racial biases.
Hassabis said Google had taken steps to address the bias in Gemini and had retrained the model. He said the new version of the image generator would be more accurate and inclusive.
A Google spokesperson confirmed Hassabis’ comments and said the company was committed to developing responsible AI technology.
“We are working to ensure that Gemini aligns with our ethical principles and will not be used for harmful or discriminatory purposes,” the spokesperson said.
The relaunch of Gemini would mark a significant step forward for Google in the field of generative AI, a type of AI that can create new content, such as images, text, and music.
Experts say the relaunch of Gemini will provide new tools for creative professionals and content creators. However, they also caution that it is important to use generative AI responsibly and to address its potential for bias and misuse.
【来源】https://www.ithome.com/0/752/224.htm
Views: 1