For more than three months, Google executives have watched Microsoft and startup OpenAI spark the imagination of internet users with ChatGPT and the AI ​​capabilities of chatbots.

Google stepped out of the spotlight today with the introduction of its own artificial intelligence chatbot called Bard. The new chatbot will initially be available to a limited number of users in the US and UK, and will gradually become available to users from other countries with more language options.

This is the company’s first move in response to the emergence of chatbots led by OpenAI and Microsoft. Google, which wants to prove that it has similar technological capabilities, introduced Bard as an independent service rather than part of a search engine.

Like other chatbots, Bard is based on a large language model, an artificial intelligence technology that learns new skills by analyzing huge amounts of data from the internet. This means that the chatbot can often be wrong about the facts, and sometimes make up information that has nothing to do with reality. a phenomenon that artificial intelligence researchers call “hallucinations”..

Google stressed that it had taken the necessary steps to curb this behavior, but acknowledged that its controls were not entirely effective.

Artificial intelligence said no

When executives demoed the chatbot on Monday, the system refused to answer a medical question because it would require accurate and correct information. The bot also provided false information about a source for an answer about the American Revolution.

In a disclaimer under the Bard’s question box, Google warns users of potential issues that may arise: “Bard may display information that is inaccurate or offensive and does not represent the views of Google.” The company also provides users with three answer options for each question and allows them to take turns providing feedback on the usefulness of Bard’s answers.

Google has been experimenting with Bard’s technology since 2015, but has only allowed access to a small group of “testers” so far because, like OpenAI and Microsoft’s chatbots, it doesn’t always provide reliable information and can still be biased towards women or women. . minorities.

“We know the problems very well – we have to responsibly attribute them to the market,” said Eli Collins, Google’s vice president of research. Collins and Sissy Xiao, Google’s vice president of product, said in an interview that the company has yet to decide how it will monetize Bard.

Artificial intelligence in applications

Google announced last week that artificial intelligence will be built into productivity apps like Docs and Sheets, which businesses pay to use. The underlying technology will also be up for sale to companies and software developers who want to build their own chatbots or launch new applications.

“This technology is still in its infancy,” Xiao said. “We’re looking into how these experiences can be linked to different products.”

The recent announcements are the start of Google’s plan to roll out more than 20 AI products and features, according to The New York Times.

Instead of being bundled with a search engine, Bard is a standalone website with a question box. At the bottom of each answer, there is a “Google” button that takes users to a new tab with the usual Google search results for the given query.

Google executives introduced Bard as a creative tool for writing emails, poetry, and counseling parents on topics such as childhood hobbies. The company wants to see how this technology will be used to further improve the chatbot based on usage and feedback.

“We’re looking at Bard as a complement to Google search,” Xiao said. “We want to be bold in how we innovate with this technology, as well as being responsible.”

Source: New York Times.