CNN
—
Google is moving forward with plans to bring AI chat features to its core search engine as it works to keep pace with a wave of new artificial intelligence tools that could threaten the company’s dominance online for the first time in decades.
The company on Wednesday said it is introducing the next evolution of Google Search, which will use an AI-powered chatbot to answer questions “you never thought Search could answer” and to help get users the information they want quicker than ever.
With the update, the look and feel of Google Search results will be noticeably different. When users type a query into the main search bar, they will automatically see a pop-up with an AI-generated response in addition to displaying traditional results.
Users can now sign up for the new Google Search, which will first launch in the United States, via the Google app or Chrome’s desktop browser. A limited amount of users will have access to it in the coming weeks, according to the company.
The updates were unveiled at I/O, the company’s annual developer event, which focused on a mix of AI and hardware products. At the event, Google also announced PaLM 2, its latest AI language model to rival ChatGPT-creator OpenAI’s GPT-4. The move marks a big step forward for the technology that powers the company’s AI products and promises to be better at logic, common sense reasoning and mathematics. It can also generate specialized code in different programming languages.
The moves come as Google’s rivals, including Microsoft, are racing to develop and deploy AI features in search engines and productivity tools in the wake of the viral success of ChatGPT. The immense attention on ChatGPT reportedly prompted Google’s management to declare a “code red” situation for its search business.
In addition to the changes to search, Google is expanding access to its existing chatbot Bard, which operates outside the search engine and can help users do tasks such as outline and write essay drafts, plan a friend’s baby shower, and get lunch ideas based on what’s in the fridge. The tool, which was previously available to early users via a waitlist only in the US, will soon be available for all users in 120 countries and 40 languages.
Google is also launching extensions for Bard from its own services, such as Gmail, Sheets and Docs, allowing users to ask questions and collaborate with the chatbot within the apps they’re using.
But incorporating AI chatbots comes with some risks. These tools have sparked concerns about tone and accuracy, the latter of which is of particular important for the online search engine that has long been the cornerstone of Google’s business.
In CNN’s virtual demo of the tool ahead of Wednesday’s announcement, the AI search tool answered questions about why bees were so important to our ecosystem, whether the Sound Hotel in Portland, Oregon has Peloton bikes (it does) and what are some local chess camps for kids, in a matter of seconds.
The tool scans websites, extracts related information and packages it neatly at the top of the results page, highlighting the sources in a section set off to the side.
But it’s not perfect; in one search for the “best pizza places in New York City,” the results were populated with restaurants from San Francisco.
Cathy Edwards, Google’s VP of search, told CNN that it’s still “very early on” and the company will continue to make changes in the weeks and months ahead.
“We really want to learn and … and iron out the kinks,” Edwards said. “We don’t want to bring this experience to everyone until we have confidence that we’ve nailed it.”
Unlike other chatbots, such as ChatGPT, Snapchat’s My AI tool and Bard, Google’s search tool is purposefully devoid of a “persona.”
“We made a deliberate decision for it to only reflect information on the web,” Edwards said. “It won’t respond with ‘I think’ or express opinions on things. It is not something that feels like a lot of other chatbots out there.”
But that choice can make for a jarring experience, if you’ve spent months using other tools. When CNN asked Google’s tool for suggestions on how to balance work and life with children at home, it offered no words of empathy or connection for the daily juggle, unlike other chatbots.
The new Google Search also offers a Perspectives feature to showcase what other people are buying or thinking about, and factoring that into results. Another tool called, About This Image, understands facts about an image, so users can ask questions about when Google first saw the picture and if it appears on other websites. The feature aims to provide “a level of understanding of an image rather than taking it at face value,” Edwards said.
These efforts highlight Google’s commitment to move forward with AI even as the technology behind it has sparked concerns.
In March, Google was called out after a demo of Bard provided an inaccurate response to a question about a telescope. Shares of Google’s parent company Alphabet fell 7.7% that day, wiping $100 billion off its market value.
Microsoft’s AI chatbot was also called out for errors made in a demo.
Like ChatGPT, the new Google Search and Bard are built on a large language model. They are trained on vast troves of data online in order to generate compelling responses to user prompts, but these tools are also known to get responses wrong or “hallucinate” answers.
Google previously told CNN that Bard would serve as a separate, complementary experience to Google Search, and planned to “thoughtfully” add large language models to search “in a deeper way” at a later time.
“We’ve been on a 25 year journey for search, and it still remains such an unsolved problem,” Edwards said. “The next long arc which will be measured in decades will be this, so we want to be bold but want to be responsible and get it right.”
