Google on February 6 unveiled “Bard” – an AI chatbot that draws on information from the web to provide high-quality responses. Google’s new conversation AI service Bard is in direct competition with ChatGPT, the AI chatbot funded by Microsoft that became an internet sensation upon its release. Like ChatGPT, Bard too will be capable of holding conversations like a human and distilling information from the web to provide coherent, concise responses.
Here is what we know so far about Bard, Google’s answer to ChatGPT:
Bard is an experimental conversational AI service, powered by LaMDA (Language Model for Dialogue Applications), a large language model that was developed and released by Google in 2021. Interesting aside: Last year, a Google employee was fired after claiming that the company’s chatbot generator LaMDA had become sentient.
In a blog post announcing the launch of Bard, Google CEO Sundar Pichai said the chatbot “draws on information from the web to provide fresh, high-quality responses.” Bard seeks to combine the breadth of the world's knowledge with the power, intelligence, and creativity of our large language models, the CEO of Google said.
People can use Bard to simplify complex topics, like explaining new discoveries from NASA’s James Webb Space Telescope to a 9-year-old. It can also help with queries like “compare two Oscar movies” or “plan a friend’s baby shower” – basically taking on complex questions and analysing different perspectives to generate a response.
The service will be initially opened up to "trusted external testers" before making it more widely available to the public in the coming weeks, Pichai said in the blogpost.
Pichai said they will be initially releasing Bard on a "lightweight" version of LaMDA that would require significantly less computing power, enabling them to scale to more users and seek more feedback.
Google will combine external feedback with internal testing to ensure Bard’s responses meet “a high bar for quality, safety and groundedness in real-world information.” This is important because chatbots learn by analysing text on the internet and are incapable of distinguishing between fact and fiction. As such, they may generate responses which are biased, racist or sexist.
In fact, it was concerns about the chatbot generating hate speech or toxic statements that had initially made Google wary of releasing the technology, the New York Times reported. The debut of ChatGPT, however, forced Google to expedite the release of their own AI services, including Bard.
Pichai told employees on Monday that the company would need all hands on deck for testing Bard. “Next week, we’ll be enlisting every Googler to help shape Bard and contribute through a special company-wide dogfood,” Pichai told employees in an email shortly after the announcement of Bard’s launch.
Where Bard could possibly have an edge over ChatGPT is answering questions on recent events, as ChatGPT’s information is restricted to internet data posted till 2021. Bard, on the other hand, is capable of generating text on recent events – like NASA’s James Webb Telescope, for example. While Bard will answer questions about the telescope’s discoveries, trying the same question with ChatGPT generates the following response: “The James Webb Space Telescope (JWST) is a highly anticipated space observatory set to launch in 2021, which will be the successor to the Hubble Space Telescope… Since the JWST has not yet launched, no new discoveries have been made using the telescope.”
More details about Bard and Google’s other AI services may be revealed at a Google event on February 8.