Google Launches Dreambeans in India to Make AI-Generated Personal Narratives Accessible to All
Google has introduced its experimental AI application, Dreambeans, in India, marking its first launch outside the United States. The company aims to leverage India’s vast internet user base to deliver personalized visual narratives based on users’ digital interactions.
On Monday, Google rolled out Dreambeans, compatible with both Android and iOS, in India, offering support for 13 regional languages, including Hindi, Punjabi, Tamil, Telugu, and Bengali, in addition to English. This launch follows just a couple of months after the app became accessible to eligible U.S. users aged 18 and older.
Launched in June, Dreambeans utilizes data from various Google services such as Calendar, Gemini, Gmail, Photos, Search, and YouTube to automatically create personalized illustrated stories that reflect users’ interests and daily lives, requiring no prompts from them. Initially, the app provided static AI-generated illustrations, but a recent update introduced animations using Google’s Veo video-generation model in certain stories.
Google opted for India as Dreambeans’ first global market due to significant interest from Indian users eager to access the app and the extensive use of Google services in the region. According to Google product manager Gozde Oznur, the country’s unique blend of users across platforms like Gemini, Gmail, and Google Photos made it an ideal choice.
Since its U.S. debut, Dreambeans has achieved nearly 55,000 downloads, as per estimates from a market analysis firm. Around 63% of these downloads occurred in September when Google made the app available to all eligible U.S. users over 18. However, there was an 84% drop in downloads in October, bringing the total down to approximately 5,500.
In its first month of availability to all eligible U.S. users, Dreambeans garnered about 29,000 downloads. In comparison, Meta’s Muse, OpenAI’s ChatGPT, and others amassed significantly higher numbers during their launch windows. However, Dreambeans differentiates itself by focusing specifically on generating personalized stories rather than providing responses to a wide range of user inquiries.
While Dreambeans accommodates several widely spoken languages in India, users can only select one language at a time for their stories. Although the app can access information in multiple languages, it currently lacks the ability to generate stories that combine languages, which may limit its appeal in a multilingual country where code-switching is common.
The app’s reliance on users’ Google account data also raises privacy concerns. Employing Google’s Personal Intelligence technology, Dreambeans accesses linked services to personalize its stories. Such services may contain sensitive information that some users might be hesitant to share with an AI application for personalized content.
Users have the option to connect specific Google services to Dreambeans and can disconnect them whenever they wish. They can also delete their data from the app, with assurances that information removed from the connected services is also deleted from Dreambeans’ personalization system.
According to Oznur, users can instruct Dreambeans to avoid certain topics, such as business-related subjects, and the app will attempt to tailor its stories accordingly. Additionally, Dreambeans is available exclusively to personal Gmail accounts and not for enterprise users, with generated stories accessible only to the respective account holder.
Dreambeans’ proactive storytelling approach carries the possibility of revisiting memories that may be distressing for some users. Oznur acknowledged this potential, mentioning that the app could produce stories about deceased individuals. Users have the option to direct the app to refrain from displaying specific topics, similar to features in Google Photos that provide memory management.
Oznur indicated that Google is actively seeking ways to enhance Dreambeans, looking to evolve beyond personalized storytelling and recommendations to enable users to act upon suggestions made by the application.
