Project Ellman: Google’s Vision for a Personalized Life Story Narrator through Chatbot Innovation
According to a recent CNBC report, Google is exploring the concept of a chatbot that delves into the narrative of users’ lives, leveraging their photo collections and search histories. The tech giant aims to integrate extensive language models like the newly unveiled multimodal Gemini model into an upcoming AI initiative. This Gemini model is positioned as a competitor to OpenAI’s GPT-4 and purportedly outperforms its closest rival on specific benchmarks, as claimed by Google.
Citing internal documents, CNBC discloses that a Google AI team proposed the development of AI-driven technology harnessing smartphone data—specifically users’ photos and search activities—to fuel an AI-based chatbot. Referred to as Project Ellman, this initiative intends to empower the chatbot with the capability to address previously unanswerable inquiries, as detailed in the report.
Contrary to relying solely on image pixels and associated metadata, Project Ellman aims to analyze patterns within a user’s photo collection. This involves examining contextual cues surrounding the images, such as the moments before and after the captured snapshot, to gain a comprehensive understanding, as outlined in the report. Additionally, Google’s internal documentation envisions “Ellman Chat” transforming into a personalized “Your Life Story Teller,” indicating its potential to craft personalized narratives from users’ accumulated data.
