Benchmarking ChatGPT, Google CEO: There will definitely be mistakes when using Bard, user feedback is welcome

On Tuesday morning, Google opened a beta version of Bard based on the large language model LaMDA, which can provide answers to various questions. After the news was announced, the stock price of Alphabet, Google's parent company, rose nearly 4% in intraday trading.

News on March 22, local time on Tuesday, Google began a public test of its artificial intelligence chat robot Bard among British and American users. "As more and more people start using Bard and testing features, they will surprise us, for sure," CEO Sundar Pichai said in an email to all employees that day. Errors will occur." "But user feedback is crucial to improving the product and underlying technology."

On Tuesday morning, Google opened a beta version of Bard based on the large language model LaMDA, which can provide answers to various questions. After the news was announced, the stock price of Alphabet, Google's parent company, rose nearly 4% in intraday trading.

But Google warns that Bard may make mistakes or "give inaccurate or inappropriate responses."

In recent months, due to the rapid development of artificial intelligence technologies such as OpenAI and ChatGPT supported by Microsoft, Google wants to keep up with the fast pace of the artificial intelligence field by releasing Bard.

When Google first announced the launch of Bard this year, many employees and investors believed that the company was acting too hastily in response to Microsoft's move to integrate ChatGPT. At a recent all-hands meeting held by the company, employees were most concerned about the purpose of releasing Bard, and some were confused about it. Google executives at the meeting argued that Bard was an experiment to differentiate the AI ​​chatbot from its core search product.

Last month, Pichai issued an appeal to all Google employees to mobilize everyone to participate in testing the Bard project, requiring employees to rewrite incorrect answers generated by the chatbot. Pichai said in an email on Tuesday that more than 80,000 Google employees are involved in testing Bard.

Pichai also said in Tuesday’s email that the company is responsible for the testing of Bard and has invited 10,000 trusted testers “from different fields and with different perspectives” to participate in the Bard project. Pichai said Google employees “should be proud of this work and the technological breakthroughs that have led us to this point over the years, including our Transformer research in 2017 and Palm and BERT. and other basic models." He added: "Even with these advances, we are still in the early stages of a long artificial intelligence journey."

I'm excited to see Bard now, Pichai said. How to stimulate user creativity and curiosity." He added that he was looking forward to sharing "a series of progress in the field of artificial intelligence" at Google's annual developer conference in May this year.

The full text of the memo sent by Pichai is as follows:

"Googlers:

Last week was an important week for artificial intelligence. We announced A series of announcements about the introduction of artificial intelligence into Google Cloud, developer platform Developer and collaboration tool Workspace. There will be more news this week as we begin to gradually open access to Bard, first announced in February.

Starting today, users in the United States and the United Kingdom can register through bard.google.com. This is just the first step, and we will continue to open Bard to more countries and regions as time goes by.

I'm grateful to the Bard team. They've probably spent more time with Bard than anyone else over the past few weeks. And a huge thank you to the 80,000 Googlers who help test Bard. We should be proud of this work. , are proud of the technological breakthroughs that have led us to this point over the years, including our Transformer research in 2017 and foundational models like Palm and BERT.

Even with these advances, we are still in a long It’s early days in the AI ​​journey. As more people start using Bard and testing features, they’ll surprise us, and things will definitely go wrong. But user feedback is critical to improving the product and underlying technology.

We have adopted a responsible development approach and invited 10,000 trusted testers from different fields and with different perspectives. We still welcome all kinds of feedback, and we will learn from it and continue to iterate and improve.

Now, I’m excited to see how Bard inspires users’ creativity and curiosity. I look forward to sharing with you a series of progress in the field of artificial intelligence at the developer conference in May this year to help people, Enterprises and communities." (Chenchen)

The above is the detailed content of Benchmarking ChatGPT, Google CEO: There will definitely be mistakes when using Bard, user feedback is welcome. For more information, please follow other related articles on the PHP Chinese website!