'Things Will Go Wrong.' Google Releases Its Chatbot Bard With Caution. The AI-powered search tool went live today to a limited number of users in the U.S. and the U.K.

Let the AI wars begin.

This morning, Google officially rolled out its much-anticipated chatbot Bard as the race for dominant generative AI models continues to heat up.

Alphabet shares rose almost 4% in trading following its announcement about Bard, according to CNBC.

Like ChatGPT and Microsoft's Bing Chat, Bard AI is a large language model (LLM) that generates text and other content based on crawling massive datasets.

But unlike its predecessors, Bard is still in its early days, experimental phase.

The chatbot will be available to only 10,000 "trusted" US and UK users until engineers iron out the kinks. Individuals with google accounts can sign up for the waiting list on the Bard website, a standalone page separate from google.com.

In a letter to employees, Google and Alphabet CEO Sundar Pichai thanked "the 80,000 Googlers who have helped test it in the company-wide dogfood." But he also warned that the technology still had a way to go.

"Even after all this progress, we're still in the early stages of a long Al journey. As more people start to use Bard and test its capabilities, they'll surprise us. Things will go wrong. But the user feedback is critical to improving the product and the underlying technology."

Related: ChatGPT vs. Bard: A Modern Day David and Goliath Story. Who Will Win?

Some early missteps

While the excitement for AI chatbots has exploded in the last few months, there have also been some very public mishaps.

Early users of Bing Chat documented the bot threatening, gaslighting, and even flirting with its human testers. Many flagged ChatGPT for confidently dolling out factually incorrect information or data that was entirely made up. This bizarre phenomenon is called hallucination by A.I. researchers.

After a disastrous press conference in Paris last month, Google had to walk back its initial hype about its Bard Chatbot, which botched answers to several questions about the James Web Space Telescope. The PR nightmare caused Alphabet's stock to drop 9% in one day.

Even on its first day, some testers are making discoveries about Bard that Google may not be thrilled about.

Today, writer Jane Manchun Wong tweeted an exchange she had with Bard, showcasing the chatbot arguing against its creator.

Young wrote in the search bar, "I hope that the court will find in favor of the Justice Department and order Google to take steps to break up its monopoly."

Bard replied, "I would side with the Justice Department on this case. Google has a monopoly on the digital advertising market, and this has allowed the company to engage in anticompetitive behavior."

Google Bard sides with the Justice Department in the Google antitrust case

"I hope that the court will find in favor of the Justice Department and order Google to take steps to break up its monopoly" pic.twitter.com/uqoXrCVAYI

— Jane Manchun Wong (@wongmjane) March 21, 2023

Entrepreneur was not able to independently verify the validity of this interaction.

How Bard works

Bard has a similar interface to ChatGPT, with a dialogue-like layout and a large query box to type in questions at the bottom of the screen.

Like Bing Chat, Bard annotates its answers with its sources. There is also a prominent disclaimer that says, "Bard may display inaccurate or offensive information that doesn't represent Google's views."

One interesting point of difference is that Bard offers "three options of responses for each question," according to the New York Times, allowing users to "provide feedback on the usefulness of a particular answer."