Today, the general public is getting access to Bard, an early experiment enabling collaboration with generative AI. This follows Google’s announcement last week that it will continue to deliver AI experiences that benefit people, businesses, and communities.
Bard claims to boost productivity, accelerate ideas, and spark curiosity. You can ask Bard for tips that will help reach your goals like reading more, planning a workout schedule, planning a trip, explaining complex topics in simple terms or drafting a blog post that inspires you. So far, Google has learned a lot from testing Bard. But now it’s time for more public testing.
Bard is based on the Research Large Language Model (LLM), specifically a version of LaMDA, and will be adjusted with newer and more powerful models over time. When prompted, Bard selects one word at a time from the possible next words to generate an answer. Since this technology is in its infancy, it’s not perfect. For example, it may appear in your results because it learns from different information that reflects real-world biases and stereotypes. It may also provide inaccurate, misleading, or false information. For example, when asked to suggest a simple houseplant, Bard came up with the wrong scientific name but got other things right.
When you use Bard, you often have several different drafts to choose from, so you can pick the best starting point for you. So, if you don’t like the first answer you get, you can always curate your question or simply ask again in order to get something that might give you that creative spark that we all need when we’re in a rut.
Work at Bard builds on AI principles at Google with a continued focus on quality and security. They use human feedback and reviews to improve their system, and they also have built-in guardrails like limiting interactions so that things stay on track. So, we can be sure that there is at least someone real at the other end…at least for now.
You can sign up over at https://bard.google.com/