Microsoft integrates ChatGPT technology into Bing search and Edge browser

Microsoft has announced a new version of the Bing search engine, which is based on an updated version of the same artificial intelligence technology that underlies the ChatGPT chatbot, reports The Verge. The updated search engine will be launched together with the new functions of the Edge browser, which will also include ChatGPT capabilities and should provide a new experience for browsing web pages and searching for information on the Internet.

Microsoft CEO Satya Nadella noted at the presentation that the paradigm of web search has not changed for decades, but artificial intelligence can provide information more flexibly 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.”

The updated Bing search engine will be able to show information in different configurations. These will be both traditional search results, but next to annotations provided by artificial intelligence, and the ability to communicate directly with the Bing chatbot, asking it questions in a chat interface like ChatGPT.

Microsoft demonstrated several search examples: asking Bing for recipes, travel tips, and buying 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 chatbot answered this question in full, describing an approximate route and providing links to sources for more information.

Unlike ChatGPT, the new Bing can also receive news about recent events. In a demo, the search engine was even able to answer questions about its own launch by referencing stories published by news websites in the last hour.

Microsoft says all of these features are powered by an updated version of GPT 3.5, the OpenAI artificial intelligence language model that underlies ChatGPT. Microsoft calls it the “Prometheus model” and says it’s more powerful than GPT 3.5 and better responds to search queries by providing up-to-date information and annotated answers.

The new Bing is available today for a limited number of users, but for now it looks like the system can provide answering only one of several preset queries and showing the same results each time. There’s also a waiting list you can sign up for to get full access to the new Bing in the future.

In addition to the new search, Microsoft is also launching two new AI features for its Edge browser: chat and compose. They will be built into the Edge sidebar.

Chat will allow users to create a short annotation of the web page or document they’re viewing and ask questions about its content, while the compose option acts as an assistant for writing texts, from emails to social media posts, based on a few initial prompts.

The announcement of the new Bing came amid a flurry of activity in the field of artificial intelligence by Microsoft and its rival Google. Since the launch of ChatGPT last November, interest in AI-powered text generation has skyrocketed. Microsoft, which works closely with ChatGPT developer OpenAI, is looking to capitalize on the hype and has already announced how the technology will be integrated into its office software suite.

Google, meanwhile, has been caught off guard by what some call a paradigm shift in how to find information on the Internet. The launch of ChatGPT reportedly caused panic inside search giant, and long-absent founders Larry Page and Sergey Brin were called in to help deal with what could be a threat to the company’s biggest source of revenue.

Trying to get ahead of Microsoft’s announcement today, Google yesterday presented ChatGPT competitor called Bard. The company’s CEO, Sundar Pichai, described the software as an “experimental conversational AI service,” but noted that it is still being tested by a small group of users and won’t get a wider launch until the coming weeks.

However, the big question for Microsoft and Google is whether AI chatbots are a good replacement for search? How will this technology fit in with existing methods of finding information on the Internet, and what happens when it makes a mistake?

The latter question is the most important, as AI language systems such as ChatGPT have a well-documented tendency to present false information as fact. Although researchers have been warning about the problem for years, since ChatGPT’s launch, countless examples of AI-induced mistakes have surfaced online, from chatbots fabricating biographical information about real people to falsifying scientific papers and giving dangerous medical advice.

This feature of AI has already become a problem. The advent of chatbots has brought renewed attention to this problem, but Google has already used AI to summarize information from web pages for years. This has led to some high-profile mistakes, such as Google’s response to the query “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).

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!”

However, the company did not touch on some issues, including how AI search could upset the balance in the Internet ecosystem. If AI tools like the new Bing pull information from sites without users going to the source, it will eliminate the revenue stream that keeps many websites afloat. In order for the new search paradigm to be successful, some of the old monetization models must be preserved or new ones must be sought to replace them.