The Paper reporter Yang Yang

　　A day after Microsoft showed off its new version of search engine Bing, Google's press conference was "not enough surprises."

　　On the evening of February 8th, Beijing time, Google held an AI (artificial intelligence) conference in Paris to showcase the company's new progress in the field of artificial intelligence. Google Maps software, multiple search functions, etc. ushered in updates, but the much-anticipated chatbot Bard is still in beta, and the official release date has not yet been announced.

　　Just the day before, Microsoft announced the launch of a new version of the search engine Bing and the browser Edge equipped with AI functions, which attracted attention in the industry.

In contrast, Google's press conference seemed "unremarkable", and even the Google team couldn't find a demo phone on stage at the press conference.

The "Wall Street Journal" called Google's big move this time "a counterattack in the field of artificial intelligence", and it seems that this counterattack seems to be going badly.

　　As of the close of U.S. stocks on the 8th, Google (Nasdaq: GOOGL) fell 7.68% to close at $99.37 per share, and its market value evaporated by more than $100 billion.

The release date of Bard is unknown, and the "Multiple Search" function is available globally

　　For the much-anticipated chatbot Bard, Google has not revealed when the official version will be released.

But the company said it plans to use the technology in search engines to provide long text answers for search results that don't have standard answers.

　　The company said the feature will be officially rolled out when Google is confident in the quality of the answers the AI ​​provides.

　　"We are going beyond the traditional concept of search to help users understand information in new ways." Prabhakar Raghavan, senior vice president of Google, said in the press conference, "The new generative AI capabilities can help us organize complex information and opinions in search results. .”

　　In addition, Google also revealed new developments in the updated version of the "multisearch" function, which can be used on mobile devices with the smart lens (Google Lens) application and is open to users around the world.

　　Multisearch, first launched last April and opened to US users last October, supports information in a variety of formats, including text, photos and videos.

According to reports, users can select their favorite shirt photo in Google search and ask Lens where to find the same style.

In addition, users can point their phones at broken parts on their bikes and type questions like "how to fix" into Google Search.

　　Google says the combination of words and images can help Google Search handle problems it couldn't handle before.

In the coming months, Google will also expand the feature beyond mobile devices, launching a multi-search feature that supports the web version.

　　Regarding Google Maps, the company said that it is currently launching a feature that allows Google Maps users to observe a three-dimensional view of a destination, such as allowing users to directly see the interior of a restaurant.

The official promotional materials were accused of giving wrong answers

　　Bard is an experimental artificial intelligence program based on the Language Model for Applications (LaMDA).

Previously, on February 6, Pichai revealed that Bard has been opened to trusted testers and will be launched more widely to the public in the next few weeks.

　　He likens Bard to "an outlet for creativity and a launching pad for curiosity", saying it can pull information from the web and provide "fresh and high-quality" answers.

　　It is not yet known how "fresh" information Bard can provide, but this may be one of the most important differences between Bard and ChatGPT.

ChatGPT, which is currently popular all over the world, has a "time difference" with the world-ChatGPT stated that its own knowledge is as of 2021, and it cannot guarantee to answer updated information.

　　However, Bard was pointed out that there were loopholes in the answer before it was officially released.

　　At the same time as Google's press conference on the 8th, the British "Daily Telegraph" reported that Bard had an incorrect answer in an official promotional article showing how it works.

　　According to Google's official blog, a user asked Bard, "What new discoveries did the James Webb Space Telescope find that I can tell my 9-year-old child?" One of Bard's replies read, "The James Webb Space Telescope has photographed exoplanets (JWST took the very first pictures of a planet outside of our own solar system.)”

　　According to the Daily Telegraph, the first photo of an exoplanet appeared in 2004, taken by the Very Large Telescope array in Chile.

Called 2M1207b, the exoplanet is about five times the size of Jupiter and lies about 170 light-years from Earth.

The report further pointed out that the NASA telescope will not be put into use until December 2021.

　　According to NASA's official website, "2M1207b is the first directly imaged exoplanet. It was first photographed by the VLT in 2004."

　　As for the description of the James Webb Space Telescope, NASA wrote in an article on September 1, 2022, "Astronomers used the James Webb Space Telescope for the first time to directly photograph (take a direct image) a part of the outer solar system. A picture of a planet."

　　Regarding the accuracy of chatbots, artificial intelligence analysts have previously warned that there are still potential problems with generative AI (generative AI). For example, ChatGPT's responses sometimes include fictional facts.

　　Chirag Shah, a University of Washington professor who studies flaws in chatbots, previously said, "Companies too often introduce new technologies too quickly, ignore their flaws, and then try to fix them quickly ... and that can do real harm."

　　The "Daily Telegraph" even directly wrote in the report that Bard's mistake will further cause people to question the accuracy of the answers generated by search engines and artificial intelligence.

　　Although the competitor ChatGPT has also been exposed to many wrong answers, for Google, which is full of confidence, it is obviously not smooth for Bard to be "caught wrong" at the beginning of its release. Google's stock price once dived nearly 9% after the above-mentioned mistakes were exposed by the media.

In the field of racing AI, can Google launch a counterattack?

　　The "Wall Street Journal" believes that Google's announcement of the introduction of AI technology in the search engine is a counterattack in the field of artificial intelligence.

　　Since OpenAI publicly released ChatGPT in November last year, major companies have sounded the alarm.

Less than a month after ChatGPT came out, Google issued a "Code Red (Red Code)" warning.

　　According to the "New York Times" report in December 2022, since the latest version of ChatGPT was released, Sundar Pichai, CEO of Google and its parent company Alphabet, has participated in several meetings around Google's artificial intelligence strategy to discuss How to deal with the threat ChatGPT poses to its search engine business.

　　The British newspaper The Times reported in January that the company had also brought in Google co-founders Larry Page and Sergey Brin to tackle the issue.

　　Just the day before the Google press conference, Microsoft also held a press conference, announcing the launch of the search engine Bing (Bing) and the browser Edge equipped with AI technology. After that, users can directly ask questions and generate answers through search engines.

　　Compared with Google's "unremarkable" conference, the news brought by Microsoft is more exciting.

The Wall Street Journal pointed out that Microsoft hopes that this artificial intelligence upgrade can help it weaken Google's dominance in the search market, which accounts for 90% of the global search engine market.