Microsoft Germany's Chief Technology Officer, Andreas Braun, has confirmed that the much-anticipated GPT-4 will be released within a week of March 9, 2023.

The most significant development with GPT-4 is its multimodality, meaning that the AI model will be able to operate across various input types like video, images, sound, and text.

Previously, GPT-3 and GPT-3.5 only operated within one modality, text. However, Microsoft has revealed that GPT-4 can operate in at least four modalities, including images, auditory, text, and video.

According to Dr. Andreas Braun, GPT-4 will introduce multimodal models that offer entirely different possibilities, including videos. It is still unclear whether these multimodalities are specific to GPT-4 or general.

“We will introduce GPT-4 next week, there we will have multimodal models that will offer completely different possibilities – for example videos…”

Microsoft Director Business Strategy Holger Kenn explained that multimodal AI can translate text into not only images but also music and video.

Microsoft's Kosmos-1, a multimodal language model, was released at the beginning of March 2023, integrating text and images.

The model was subjected to various tests, with good results in classifying images, answering questions about image content, automated labeling of images, optical text recognition, and speech generation tasks. Visual reasoning, drawing conclusions about images without using language as an intermediate step, is a key feature of Kosmos-1.

GPT-4's multimodality goes beyond Kosmos-1 by adding a third modality, video, and incorporating sound. The AI model appears to work across all languages, receiving a question in one language and answering in another. Microsoft's confidence metrics also make the AI model more reliable by grounding it with facts.

There is no current announcement on where GPT-4 will be utilized. Azure-OpenAI was mentioned specifically, but the possibilities for GPT-4 are endless.

Google has been struggling to catch up with Microsoft's implementation of AI technology, with Microsoft seemingly ahead in the race. While Google integrates AI into multiple products, such as Google Lens and Google Maps, Microsoft's implementation is more visible.