Although the Pixel Fold and Pixel 7a devices will be the highlights of Google’s developer conference, AI will also feature prominently during the keynote on May 10th. CNBC has obtained internal documents about the keynote that reveal Google’s plan to unveil several generative AI enhancements, including PaLM 2 – the newest and most sophisticated iteration of its large language model.
This week, Google is set to unveil PaLM 2, its latest AI language model.
According to the documents, PaLM 2, which Google has been referring to internally as the “Unified Language Model,” has been tested on coding tasks, math and creative writing assessments, and more. The model is expected to support over 100 languages.
At the keynote, Google will focus on the theme of “enabling people to achieve their maximum potential” with “generative experiences” in Bard and Search. Sundar Pichai, the CEO of Google, and his team will showcase Bard’s capabilities in coding, solving math and logic problems, and expanding its chatbot features to include Korean and Japanese languages.
According to CNBC, Google is developing a multi-modal chatbot named “Multi-Bard,” which can solve more complex problems by leveraging a larger dataset. Google has also experimented with versions of the chatbot known as “Big Bard” and “Giant Bard.”
Additionally, Google is expected to enhance its “Workspace AI collaborator” by introducing template generation in Sheets and image generation in Slides and Google Meet. The leaked documents featured an image that showed a chatbox in a Slides sidebar, enabling users to enter text and create an image based on the given words.
