In a statement made by Sundar Pichai (Google and Alphabet CEO) on February 6, Google, in an effort to be more useful to its users, is launching its own AI-based chat service, Bard. It will work much similar to ChatGPT and is designed to simulate conversations with a human using a combination of NLP and machine learning to provide realistic, helpful responses to user queries. But unlike ChatGPT, Bard is going to draw in all its information from the internet.
It’s no surprise, however, that Google chooses to launch its own AI chat tool at this point in time. The past few months talked a lot about AI. Right from its launch, ChatGPT has stirred curiosity among users, and though the underlying technology of ChatGPT is not new, users at large have, for the first time, been introduced to something like this. And Google wants to be in the hot run of AI chat tools. Bard can be integrated into message platforms, websites, desktop and mobile apps, and a range of digital systems. Google plans to launch Bard soon enough for the general audience. For now, it is in a beta test run.
What is Google Bard?
Google Bard is an experimental AI chat service by Google that was announced on February 6 by Sundar Pichai, the CEO of Google and Alphabet. It operates in a conversational manner, similar to ChatGPT, but with a key difference: it retrieves its information from the web. Despite being a new concept at its unveiling, Bard is powered by Google’s Language Model for Dialogue Applications (LaMDA), which was introduced two years prior.
How Does Google Bard Work?
Google Bard uses Google’s Language Model for Dialogue Applications (LaMDA), built on Transformer, a neural network architecture that Google invented and open-sourced in 2017. Interestingly, according to Google, GPT-3, the language model used by ChatGPT, also uses Transformer. The initial version of Bard uses a lightweight model version of LaMDA, which requires less computing power and can be scaled to accommodate more users. To provide responses, Bard draws on all the information available on the web, which Google CEO Sundar Pichai believes will provide fresh and high-quality responses.
What did Google Bard Get Wrong?
The launch of Google Bard was marred by controversy after a demo of the AI chat service delivered inaccurate information about the James Webb Space Telescope (JWST). The demo, which Google tweeted, prompted users to ask, “What new discoveries from the James Webb Space Telescope can I tell my 9-year-old about?” However, the response was factually incorrect, and users quickly noticed the error. In response, Google stated they are starting a rigorous testing process with their Trusted Tester program.
In addition to the Bard controversy, Google’s LaMDA, which powers Bard, also faced scrutiny before its release. Former Google engineer Blake Lemoine alleged that LaMDA might be “sentient” in a document he shared shortly after LaMDA’s publication. However, Google denied the sentience claims and Lemoine was put on paid administrative leave before letting him leave the company.
What other AI Services does Google Have?
Google has developed other AI services but has not released them to the public yet. Google usually deals with AI products lightly and does not release them until it is absolutely confident about their quality or usefulness.
For example, when released, Google’s AI image generator Imagen can be a great alternative to OpenAI’s DALL-E. The AI Music generator of Google, named MusicLM, has also been developed, but Google does not plan to release it anytime soon.