Microsoft has announced a new version of its search engine Bing, powered by an upgraded version of the same AI technology that underpins chatbot ChatGPT. The company is launching the product alongside new AI-enhanced features for its Edge browser, promising that the two will provide a new experience for browsing the web and finding information online.

“It’s a new day in search,” said Microsoft CEO Satya Nadella at an event announcing the products. Nadella argued that the paradigm for web search hasn’t changed in decades, but that AI can deliver information more fluidly and quickly than traditional methods.

“The race starts today, and we’re going to move and move fast,” Nadella said. “Most importantly, we want to have a lot of fun innovating again in search, because it’s high time.”

In demos today the company showed what it’s calling “the new Bing” working in various configurations. One of these shows traditional search results side-by-side with AI annotations (above), while another mode lets users talk directly to the Bing chatbot, asking it questions in a chat interface like ChatGPT (below).

Microsoft showed a number of example searches: querying Bing for recipes, travel tips, and shopping for furniture from Ikea. In one demo Bing was asked to “create an itinerary for each day of a 5-day trip to Mexico City.” The question was answered entirely by the chatbot, which described a rough itinerary alongside links to sources for more information.

Unlike ChatGPT, the new Bing can also retrieve news about recent events. In The Verge’s demos, the search engine was even able to answer questions about its own launch, citing stories published by news sites in the last hour.

Microsoft says these features are all powered by an upgraded version of GPT 3.5, the AI OpenAI language model that powers ChatGPT. Microsoft calls this the “Prometheus Model,” and says it’s more powerful than GPT 3.5, and better able to answer search queries with up-to-date information and annotated answers.

The new Bing is live today “for desktop limited preview,” but it appears users are only able to “ask” one of a number of preset queries and receive the same results each time. There is also a waitlist to sign up for full access in the future.

In addition to the new Bing, Microsoft is launching two new AI-enhanced features for its Edge browser: “chat” and “compose.” These will be embedded within Edge’s sidebar.

“Chat” allow users to summarize the webpage or document they’re looking at and ask questions about its contents, while “compose” acts as a writing assistant; helping to generate text, from emails to social media posts, based on a few starting prompts.

The announcement of the new Bing comes amid a flurry of AI activity from Microsoft and rival Google. Since ChatGPT launched on the web last November, interest in AI text generation has exploded. Microsoft, which has closely partnered with ChatGPT creator OpenAI, is seeking to capitalize on this excitement and has already announced how this technology will be integrated across its suite of office software.

Google, meanwhile, has been caught off guard by what some are claiming is a paradigm shift in how users find information online. The launch of ChatGPT reportedly triggered a “code red” within the search giant, with long-absent founders Larry Page and Sergey Brin drafted to help deal with what could be a threat to the firm’s biggest revenue driver.

In an attempt to preempt Microsoft’s announcement today, Google unveiled its own ChatGPT, named Bard, yesterday. CEO Sundar Pichai described the software as an “experimental conversational AI service” but noted that it was still being tested by a small group of users and will only receive a wider launch in the coming weeks.

The AI-powered future of search

The bigger question for both Microsoft and Google, though, is: are AI chatbots a good replacement for search? How will this technology sit alongside existing methods of finding information online, and what happens when it makes mistakes?

The latter point is the most important, as AI language systems like ChatGPT have a well-documented tendency to present false information as fact. Although researchers have warned about this problem for years, there have been countless examples of AI-generated errors since ChatGPT launched on the web — from chatbots making up biographical details about real people to fabricating academic papers and offering dangerous medical advice.

This sort of AI stupidity is already a problem, though. The rise of chatbots has generated new attention for the issue, but Google has been increasingly using AI to summarize web pages for years. This has led to some high-profile errors, like Google responding to a search “had a seizure now what?” with the advice “hold the person down or try to stop their movements” — exactly the opposite of what should be done in this scenario.

Microsoft referenced these and other issues in its presentation, saying it had been working hard to safeguard against risks like bias and “jailbreaking” (tricking AI chatbots into disregarding filters intended to prevent them generating dangerous or hateful content). “With this product, we have gone further than we ever have before to develop approaches to measurement to risk mitigation,” said Sarah Bird, responsible AI lead for Azure.

However, the company is also evidently preparing for its systems to make mistakes (though the company will be hoping not as badly as its failed 2016 chatbot Tay). The interface for the new Bing includes a warning to users: “Let’s learn together. Bing is powered by AI, so surprises and mistakes are possible. Make sure to check the facts, and share feedback so we can learn and improve!”

There were some issues that the company didn’t address, though, including how AI-assisted search could unbalance the web’s ecosystem. If AI tools like the new Bing scrape information from the web without users clicking through to the source, it removes the revenue stream that keeps many sites afloat. If this new paradigm for search is to be a success, it will need to keep some old agreements in place.