Bard AI - Google's rival to ChatGPT is now available for US and UK users to test

Google is releasing its Bard AI platform to a select group of users after months of public discussion about new chatbots driven by artificial intelligence.

‍

Those in the U.S. and the U.K. are legally permitted by the company. Register at bard.google.com for its generative AI product. In a blog post published on Tuesday, Google stated that it would gradually increase availability in new locales and languages. The article was written by Sissie Hsiao, product vice president, and Eli Collins, research vice president, and was headlined, "Test Bard and leave your feedback."

‍

They claimed that Bard might "increase your productivity, speed up your ideas, and feed your curiosity." "You may ask Bard to give you advice on how to accomplish your resolution to read more books this year, to simplify quantum physics, or to inspire your creativity by writing a blog post. Testing Bard has taught us a lot so far, and gathering feedback from more users will be crucial to furthering its improvement.”

‍

Testers must log in using a Google account.

‍

To prevent interactions between people and AI from getting out of hand, Google is erecting certain barriers. The firm won't restrict the number of daily talks, a spokeswoman said, but it will cap the number of conversational exchanges "in an attempt to keep interactions beneficial and on topic."

‍

In a briefing, Bard announced that it would present three different drafts of each response, confirming earlier this month's reporting from Trade Algo. Google anticipates that the trial phase will help it learn how frequently users interact with Bard.

‍

After the November release of OpenAI's ChatGPT technology, which has subsequently been incorporated into a limited version of Microsoft's search engine Bing, the debut comes after months of quick testing.

‍

In January, Trade Algo reported that Google had been testing its LaMDA technology in the Bard chatbot and on search results pages since late 2022. The following week, Google made product announcements.

‍

With Bard's initial release, which appeared hurried to compete with Microsoft's recently announced Bing integration of ChatGPT, Google received criticism from both employees and investors.

‍

Google announced last week that a select group of users of Workspace, which includes Gmail and Google's productivity tools, would get access to its AI technologies.

‍

Expect mistakes

‍

Each Bard query will have a "Google it" button that, according to a blog post published by Google on Tuesday, will open an additional tab on the company's search website. According to a representative who talked to CNBC, Google sees Bard's current format as an addition to search.

‍

The company is also trying to warn that Bard might err.

‍

Users are encouraged to use the "Google It" function to fact-check Bard's responses if they are in question and are reminded that Bard is an experiment and that it "may deliver wrong or inappropriate responses" when they first open it.

‍

Hsiao and Collins noted that they used Bard to help create the blog post and conceded "it didn't always get things right." Accuracy is currently a key concern with large language models or LLMs. The text also displays alongside results stating that the information "doesn't represent Google's opinions."

‍

Google officials asked last month that all staff members work together to correct Bard's incorrect responses.

‍

“Although LLMs are a fascinating technological advancement, there are several drawbacks,” according to Tuesday's post. For instance, because they are exposed to a wide variety of information that reflects biases and stereotypes from the real world, those elements occasionally appear in their outputs. Also, they have the capacity to confidently offer information that is untrue, misleading, or erroneous.

‍

A "thumbs up" and "thumbs down" option will be available in Bard inquiries to allow for user feedback.