Gemini's in-depth research mode, recently expanded to 40 languages, aims to provide users with an AI-powered research assistant to create structured reports from information.
HyunJeong Choe emphasized the importance of reliable sources for Gemini's multi-language support, acknowledging challenges in providing accurate summaries, particularly in languages like Hindi.
Choe stated that ensuring factuality remains a significant challenge for generative AI, highlighting ongoing efforts to refine the model's accuracy using native language data.
Jules Walter mentioned quality checks through native testing programs, while the company continues to gather data necessary for effectively training the AI models.
Collection
[
|
...
]