Microsoft’s New ChatGPT-Powered Bing Goes Live Today
Microsoft reps came out swinging Tuesday in a massive press blitz, though they were acting more like an artificial intelligence-driven Terminator than any live human being. Sure, the product folks at the company’s conference held at Microsoft’s Redmond, Washington headquarters and on the company blog talked with all too familiar nerdy excitement, but you can tell Microsoft talked up its newest ChatGPT-like integration into its long-maligned search engine with only one thought in mind: “kill Google.”
In a Tuesday blog post, Microsoft went all in on AI-driven searches. The company launched the new AI-powered Bing and Microsoft Edge browser incorporating an AI copilot and chat feature.
Last month, Microsoft announced a multi-billion dollar partnership with OpenAI, the company that created the DALL-E 2 AI image generator and the ChatGPT AI language model.
The new search engine, working on what Microsoft has dubbed OpenAI’s “Prometheus Model” should allow users to input a question and receive a detailed, up-to-date, and annotated answer along with citations from relevant sites. The new Bing search will show a list of links on the left, as usual, but will also showcase an explanatory section on the right. The company claimed this new system is “more powerful than ChatGPT.”
The new Bing search should be available on a limited preview for desktop, but the company said users can visit the Bing website to sign up for a waitlist. The previews for certain prompts are currently very limited, though the company plans to scale the preview version in the coming weeks and will also offer a mobile preview as well.
In a spitfire talk with journalists, Microsoft showed off “the new Bing,” or more specifically, how the AI-driven ChatGPT-like system will be integrated into Bing search. There’s two parts to this: the search function and the chat bot. The new interactive chat lets users refine their search after their initial prompt. The company said the new system should allow people to get a detailed itinerary of a planned vacation, or a list of the best TVs to buy alongside citations from different websites.
According to images at the press conference shown by The Verge, Microsoft used several examples such as “compare the most influential Mexican artists” which gave an explanatory list that included Frida Kahlo and Diego Rivera along with a short passage about their work. The company also showed that the AI could show dimensions of a product and guesstimate if it will fit in a space. Microsoft showed the system could cough up the dimensions of a Ikea Klippan loveseat and speculate it “might fit” inside a 2019 Honda Odyssey with all the seats down.
The new chat also has integrations along the same lines as ChatGPT, letting users ask the AI to write emails for them or create a quiz. Bing has long been the butt of jokes for how irrelevant its search results have been, but Microsoft claimed it “applied the AI model to our core Bing search ranking engine, which led to the largest jump in relevance in two decades.”
“AI will fundamentally change every software category, starting with the largest category of all – search,” Microsoft CEO Satya Nadella said in the blog post.
Microsoft Edge is also changing to include Chat directly on the Edge sidebar. The company used the example of asking the system to compose a LinkedIn post with a few prompts. Microsoft even claimed users can manipulate and further refine the tone of the content it produces. The new Edge also gives users the option to input content like a long financial report and then force it to offer you bullet points of the main takeaways.
How Will Micorosoft Deal With the Lingering Issues With AI?
Of course, let’s not forget how quickly this has all come together. The incredibly fast-paced development of generative AI systems has only now been eclipsed by how quickly companies are trying to integrate the technology into existing systems. Google beat Microsoft to the punch Monday when it announced its own AI called “Bard” is in the testing phase, but there’s no word when the Microsoft competitor’s system will be ready.
Developers have only started to scratch the surface for what issues could come from this tech. The company made a passing attempt in its blog post to say it would be working to implement “safeguards to defend against harmful content.” However, there’s so much more at stake than just disinformation, as bad as it is. Thousands of people could be generating prompts using this system, but the well-known ChatGPT system is inherently limited. It’s been proved the system can make up references and produce inaccurate information. This might be harmless when asking it to create a plan for a day trip, but what if it gives a user wrong medical advice during an emergency?
Sam Altman, OpenAI CEO, tweeted just two months ago that “it’s a mistake to be relying on [ChatGPT] for anything important right now.”
The company and Altman again tried to reassure people they were taking a considerate approach, despite the extremely short lead time. According to The Wall Street Journal, Altman told reporters at the press conference that “[Microsoft and I] share a deep sense of responsibility in ensuring that AI gets deployed safely.”