Bing might soon have the capability to use images, videos, other data types in responses

March 11, 2023

The new ChatGPT-powered Bing is continuously improving, and Microsoft might start its biggest development next week. Andreas Braun, Microsoft Germany CTO, announced recently that GPT-4 is coming next week. Alongside this, the executive hinted at “multimodal models that will offer completely different possibilities.”

Microsoft has already invested billions of dollars in its AI initiatives, especially in relentlessly improving Bing in order to better compete with Google. Now, the software giant confirmed the arrival of GPT-4 next week, which is expected to be injected into its search engine and chatbot.

Prior to the release of ChatGPT Bing, there were rumors about Bing using GPT-4. However, Microsoft has instead used the GPT-3.5 model alongside its proprietary tech Prometheus, which allows Bing to generate up-to-date data. Surprisingly, even though the new Bing is still not available to everyone, the company already has plans to give the search engine a significant boost via the upcoming GPT-4.

OpenAI’s new and upcoming Large Language Model is expected to allow the Bing chatbot to generate faster results, which can be a huge help since the current version usually takes a few seconds to start generating responses. Nonetheless, aside from speed, a multimodal capability could be the biggest thing the new LLM introduction could bring.

At Microsoft’s AI in Focus – Digital Kickoff event, Braun shared some details of what to expect from the entrance of a new LLM. (via Heise)

“We will introduce GPT-4 next week, there we will have multimodal models that will offer completely different possibilities – for example videos,” said Braun, who described LLM as a “game changer.”

Moreover, Braun confirmed that Microsoft has plans to “make the models comprehensive” using the multimodality capability. Once injected, this should allow Bing to provide a variety of data when responding to questions, which means it could also process videos, images, and other data types. This should result in better responses, making Bing a more effective search assistant for everyone.

On the other hand, it is important to note that Bing is not the first in multimodality. Recently, rolled out its multimodal chat search feature, allowing users to provide text and voice inputs and receive responses beyond conversational texts. The search engine, however, is still struggling to get attention from the public. Meanwhile, despite not yet being fully accessible by all, Bing already has a continuously expanding waitlist. The injection of a multimodal capability into it will certainly affect its rivals, such as, it is still early to say how vast this effect will be since Braun’s announcement only confirms very few details.

Leave a Reply

Your email address will not be published. Required fields are marked

{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}