Google has announced a series of updates to Bard, its experimental AI chatbot that aims to provide fresh, high-quality responses to users’ queries and requests. Bard, which is powered by Google’s Language Model for Dialogue Applications (LaMDA), can now connect to Google apps and services, such as Gmail, Docs, Drive, Maps, YouTube, and Flights, to help users with a wider variety of tasks. Google has also improved the “Google it” feature to let users easily double-check Bard’s answers, and expanded features to more places.
Bard Extensions: A new way to interact with Google tools
One of the most notable updates to Bard is the launch of Bard Extensions, a feature that allows Bard to find and show relevant information from the Google tools users use every day. With Extensions, users can ask Bard to access their Gmail, Docs, and Drive content, as well as look up information from Google Maps, YouTube, Flights, and hotels, even when the information is across multiple apps and services.
For example, if a user is planning a trip to the Grand Canyon, they can ask Bard to grab the dates that work for everyone from Gmail, look up real-time flight and hotel information, see Google Maps directions to the airport, and even watch YouTube videos of things to do there, all within one conversation. Or, if a user is applying for a new job and using Bard to help with the process, they can ask Bard to find their resume from Drive and summarize it to a short paragraph personal statement, and continue collaborating on their cover letter.
Google said that Bard Extensions are available in English, and that users can control their privacy settings when deciding how to use these extensions. Google also said that users’ content from Gmail, Docs, and Drive is not seen by human reviewers, used by Bard to show ads, or used to train the Bard model.
Google it: A better way to verify Bard’s responses
Another update to Bard is the improvement of the “Google it” feature, which lets users easily double-check Bard’s answers. When users click on the “G” icon, Bard will read the response and evaluate whether there is content across the web to substantiate it. When a statement can be evaluated, users can click the highlighted phrases and learn more about supporting or contradicting information found by Search.
Google said that this feature is available for responses in English, and that it helps users to assess the quality, safety, and groundedness of Bard’s responses. Google also said that it is working on ways to make Bard more transparent and accountable, such as providing sources and confidence scores for its answers.
More features in more places: A better way to use Bard
The third update to Bard is the expansion of features to more places, such as Google Workspace, Google Photos, and Google Assistant. Google said that users can now use Bard alongside these apps and services, and access features such as voice input, image recognition, and smart suggestions. For example, users can ask Bard to create a photo album from their Google Photos, or to help them write an email using voice input from Google Assistant.
Google said that these features are available in select languages, and that it plans to bring Bard to more platforms and devices in the future.
Bard: An experiment to collaborate with generative AI
Bard is an experimental conversational AI service that Google launched in February 2023, as part of its efforts to advance the state of the art in AI. Bard is powered by LaMDA, Google’s large language model that can generate natural and coherent responses to any kind of query or request. Bard seeks to combine the breadth of the world’s knowledge with the power, intelligence, and creativity of large language models, and to provide users with a new way to interact and collaborate with generative AI.
Google said that Bard is still in the testing phase, and that it is currently open to trusted testers ahead of making it more widely available to the public in the coming weeks. Google said that it is using external feedback and internal testing to make sure Bard’s responses meet a high bar for quality, safety, and groundedness. Google also said that it is using a lightweight version of LaMDA, which requires significantly less computing power, to scale Bard to more users and allow for more feedback.