MobileNews

Google has tasked the Assistant team to work on Bard

Google Assistant team

The conversational aspects of generative AI and large language models (LLMs), like LaMDA, make it ripe for smart assistants. Google specifically mentioned Assistant two years ago as a product where it hopes to incorporate “better conversational features.” That has yet to happen, but the Google Assistant team is helping lead Bard development.

Sissie Hsiao was appointed as vice president and general manager of Google Assistant in 2021. The longtime (starting in 2006) Google employee previously worked on Search, Docs, Analytics, and advertising.

This morning, Hsaio, along with a Google Research VP, Eli Collins, announced that access to Bard is opening up.

The company is quick to tell 9to5Google that Bard, as an experimental service, is separate from Google Assistant, but it did see the crossover and synergies between the two efforts. The Assistant team has worked on conversational AI and speech understanding for many years, and that skill set can be and has been applied to ongoing Bard development.

More importantly, the Assistant team has a lot of experience in turning years of research and development into products for end users and maintaining a large service. As of May 2022, over 700 million people every month use Assistant.

As we’ve previously written about, Google has spent recent years, 2022 especially, making Assistant more natural to use with Look and Talk on the Nest Hub Max and Quick Phrases, which are also available on Pixel phones, that let you skip the “Hey Google” hotword. More promising is how Assistant will ignore “umm,” natural pauses, and self-corrections as you’re issuing a voice command. In fact, it was Hsiao that announced these features on stage at I/O 2022.

Again, “Bard is an experiment” and is just a vehicle for Google to get training data and feedback. However, there’s clearly interest in this chat method of interacting with information, and its path to becoming a full Google product is there.



Author: Abner Li
Source: 9TO5Google

Related posts
AI & RoboticsNews

DeepSeek’s first reasoning model R1-Lite-Preview turns heads, beating OpenAI o1 performance

AI & RoboticsNews

Snowflake beats Databricks to integrating Claude 3.5 directly

AI & RoboticsNews

OpenScholar: The open-source A.I. that’s outperforming GPT-4o in scientific research

DefenseNews

US Army fires Precision Strike Missile in salvo shot for first time

Sign up for our Newsletter and
stay informed!