Google has released a new ChatGPT-like AI service called Bard to a select set of testers / Cesc Maymo. Google said it is rolling out new sea...
 |
| Google has released a new ChatGPT-like AI service called Bard to a select set of testers / Cesc Maymo. |
The moves, announced at an event Wednesday in the French capital, come a day after Microsoft Corp. said it was building the technology behind the chatbot ChatGPT into its Bing search engine, and amid a scramble by companies to roll out tools that use a type of artificial intelligence that can generate content—from haikus to high-resolution images.
The Alphabet Inc. unit said Wednesday that it is planning to build this technology like that in ChatGPT into search results to give more lengthy textual responses to complex queries with no single correct answer—such as what are the best constellations to look at. That feature will roll out when Google is confident in the quality of the answers the AI provides, the company said.
“We’re moving beyond the traditional notion of search to help you make sense of information in new ways,” Prabhakar Raghavan, a Google senior vice president, said at the event. “New generative AI features will help us organize complex information and multiple viewpoints right in Search.”
Another feature Google announced Wednesday will expand users’ ability to query the search engine based on images and videos they are looking at on their phones if they use Google’s Android operating system, allowing a user to identify, say, a local landmark.
The company also said it is now rolling out a feature that allows Google Maps users to explore three-dimensional representations of destinations—like the inside of a restaurant—extrapolated by AI from ordinary two-dimensional photos. And it said it is broadening the availability of a feature that lets users search maps for local businesses by pointing their phone at the nearby area.
“We have a lot of hard and exciting work ahead to build these technologies into our products and continue bringing the best of Google AI to improve people’s lives,” Alphabet Chief Executive Sundar Pichai said in an internal company email on Monday, which was viewed by The Wall Street Journal.
Google’s announcement of AI enhancements to its search engine is its latest counteroffensive in a tit-for-tat AI battle. Microsoft said Tuesday that it would integrate ChatGPT technology into its Bing search engine, allowing users to pose questions in natural language and receive direct responses. Microsoft said last month it is making a multiyear, multibillion-dollar investment in OpenAI, maker of ChatGPT and other AI tools such as the Dall-E 2 image-generation software.
Google said Monday it was rolling out a new ChatGPT-like AI service called Bard to a select set of testers, with a broader public launch in coming weeks. The new experimental service generates textual responses to questions posed by users, based on information drawn from the web. Bard, unlike the initial version of ChatGPT, can access information from the real world through Google Search, according to a screenshot of a response by Bard viewed by the Journal.
Those developments are part of a fast-spreading AI war over the commercial potential of so-called generative AI—artificial intelligence that can create content in response to short user inputs—since OpenAI moved to release ChatGPT publicly late last year. Microsoft has promised to integrate capabilities from generative AI tools from OpenAI across all of its products quickly, as well as making them available to outside developers.
Others are jumping into the fray as well. China’s Baidu Inc. is developing an AI-powered chatbot similar to ChatGPT called “Ernie bot,” which it plans to launch next month. The race is forcing Google—by far the most widely used search engine—into an unusual position of playing catch up.
In part, according to Google executives, that is because the company has been reluctant to roll out tools that, like ChatGPT, can sometimes spout false information or nonsense in response to user queries. The company has also been under scrutiny by researchers, regulators and its own staffers to police its own use of AI.
Google’s new Bard system appeared to fall victim to that pitfall on Monday when an example the company posted of its responses claimed that the James Webb Space Telescope took “the very first pictures” of an exoplanet outside the Solar System. The National Aeronautics and Space Administration says on its website that the first images of an exoplanet were taken as early as 2004 by a different telescope.
“It’s a good example [of] the need for rigorous testing,” Elizabeth Reid, Google’s vice president and general manager of search, said Wednesday, adding that the example was nuanced. The company declined to give a precise date for Bard’s public launch, but said the company is conducting tests to make sure it delivers accurate information. “The real bottleneck for us is to get to a place where we can get the quality where we want it to be,” Mr. Raghavan said.
Google on Wednesday emphasized, as it has in the past, that it already bakes AI into its search results, for instance by helping the search engine understand natural-language queries, even if it responds via links and snippets.
In 2018, Google created a series of AI principles that it said it would apply to its work going forward. Those rules include requirements that the AI tools should be socially beneficial, they should avoid reinforcing biases and they should be built and tested for safety in constrained environments.
Researchers cite many examples of the potential dangers. AI technology known as deep fakes, for instance, can create video that appears to be of real people saying or doing things they never said or did.
“We’ve been focusing on responsible AI since the very beginning,” Mr. Raghavan said at the event, adding that the company is “committed to setting the highest standard on how to bring it to people in a way that’s both bold and responsible.”