Google Bard is the company’s response to ChatGPT

The success of ChatGPT has been overwhelming. The AI-powered chatbot already has 100 million users, making it one of the fastest-growing services ever, if not the fastest. It was only a matter of time before Google threw its hat into the ring on this one.

Today, the company officially unveiled Google Bard. This is also an AI-powered chat service very similar to ChatGPT. However, this system is powered by LaMDA — which stands for “Language Model for Dialogue Applications.” Google unveiled LaMDA at Google I/O 2021.

Specifically, Bard is using a smaller, lightweight version of the full LaMDA tool. However, the result is the same: a human inputs a request in a conversational style, and Bard returns back an intelligent response using information drawn from the internet. The example Google uses in its announcement post is asking for simple explanations about the James Webb Space Telescope for a father to tell his 9-year-old child.

Today, Google is opening Bard up to a select group of “trusted testers.” Google would like to push out a public version in the coming weeks, but wasn’t more specific than that.

Google Bard: Why Google is holding back

The biggest problem with ChatGPT is that it can sometimes return misleading, wholly inaccurate, and even offensive results. This is because the system pulls information from all over the internet and can’t easily discern what’s true/false or appropriate/inappropriate.

This is the primary reason Google hasn’t launched a LaMDA-powered product to the public yet. The people behind ChatGPT didn’t have as much to lose as Google, which is why they launched their product without concern for the ethical and financial ramifications of such a system. Now that ChatGPT is in the wild and a massive success, Google is in a corner — it simply had to launch something.

In Google’s announcement, Sundar Pichai very deliberately lays out that Google Bard could respond with inaccurate or inappropriate information. The company implemented a system that allows users to rate responses from Bard with either a like, a dislike, or a “Check it,” which looks like a way to validate the information presented in the response.

Pichai, in the blog post, says, “We’ll combine external feedback with our own internal testing to make sure Bard’s responses meet a high bar for quality, safety, and groundedness in real-world information.”

We’ll be sure to let you know when Bard becomes available to the public.