谷歌已将其实验性人工智能应用 Dreambeans 正式拓展至印度,这是该工具在美国以外发布的首个市场 1。此次推广旨在通过支持包括印地语、旁遮普语、泰米尔语、泰卢固语和孟加拉语在内的 13 种当地语言,向大众普及由 AI 生成的个性化视觉故事 1。Dreambeans 的核心功能是利用用户连接的谷歌服务数据,具体涵盖日历(Calendar)、Gemini、Gmail、照片(Photos)、搜索(Search)以及 YouTube,从而主动生成基于个人经历的叙事内容 1。
尽管被视为拓展全球受众的关键一步,Dreambeans 此前在美国市场的表现相对温和 1。根据 Sensor Tower 的数据,自今年 6 月上线以来,该应用累计下载量接近 5.5 万次,其中约 63% 的下载发生在 9 月向所有符合条件的美国用户全面开放之后 1。在首 30 天内,其美国下载量估计为 2.9 万次,这一数字远低于 Meta 的 Muse(740 万次)和 ChatGPT(280 万次)等通用型 AI 应用 1。此外,该产品目前仍面临多语言混合输入的限制,用户仅能选择单一语言生成故事,暂不支持如“Hinglish”这样的混合语种 1。针对隐私争议,谷歌产品经理 Gozde Oznur 指出,若用户删除底层连接服务中的信息,这些数据也会从 Dreambeans 的个人化系统中移除 1。
Google has launched its experimental AI application, Dreambeans, in India, marking the app’s first release outside the United States 1. The service utilizes data from users’ connected Google services—including Calendar, Gmail, Photos, Search, YouTube, and Gemini—to proactively generate personalized visual stories 1. This move highlights Google’s view of India as a key market for expanding its global audience 1.
To support local preferences, Dreambeans is available in 13 regional languages, including Hindi, Punjabi, Tamil, Telugu, and Bengali 1. However, the current version restricts users to selecting only one language for story generation, with no support for mixed-language inputs such as Hinglish 1. Regarding privacy concerns, Google Product Manager Gozde Oznur stated that deleting information from the underlying connected services also removes it from Dreambeans’ personalization system 1.
According to data from Sensor Tower, Dreambeans has recorded nearly 55,000 downloads since its U.S. launch in June 1. Approximately 63% of these downloads occurred after the app opened access to all eligible American users in September 1. Despite this growth, the estimated download volume for the initial 30 days in the U.S. was 29,000, which is significantly lower than general-purpose AI apps such as Meta’s Muse (7.4 million) and ChatGPT (2.8 million) during comparable periods 1.
评论
还没有评论,欢迎留下第一条。