AI Models Unveiled by Microsoft for the First Time
Microsoft has announced the introduction of two new AI models - MAI-Voice-1 AI and MAI-1-Preview - as part of its growing AI division. This move comes amidst reports of OpenAI's potential restructuring from a non-profit to a for-profit company, signalling a shift towards greater independence.
The contract between Microsoft and OpenAI, which has been instrumental in powering Microsoft's Copilot AI assistant, officially expires in 2030. However, recent developments suggest a complex interplay between collaboration and competitive positioning.
In 2023, Microsoft collaborated closely with OpenAI by integrating OpenAI's AI technology into its own products, accelerating its AI capabilities faster than competitors. Yet, during the same period, Microsoft began developing its own AI models, such as MAI-Voice-1 and MAI-1-Preview, which compete directly with OpenAI's solutions.
MAI-Voice-1, described as Microsoft's first "expressive and natural speech generation model," can generate a minute of audio in less than a second using one GPU. It delivers high-resolution, expressive audio quality in scenarios with one or multiple speakers. Currently, MAI-Voice-1 is being used in the "Copilot Daily" feature and podcasts.
MAI-1-Preview, on the other hand, is the first "end-to-end trained base model of MAI," trained on around 15,000 Nvidia H100 GPUs. It is designed to follow instructions and provide helpful answers to everyday questions. MAI-1-Preview is now available for public testing on the LMArena platform.
Microsoft AI believes coordinating a range of specialized models for different user intentions and application cases will create added value. In line with this, plans are underway to integrate MAI-1-Preview into certain text application cases in its Copilot AI assistant.
Interested users can try out MAI-Voice-1 in Copilot Labs, where they can input what they want the AI to say and change the voice and speech style. OpenAI, meanwhile, is reportedly having disputes with Microsoft over changes in Microsoft's stake in OpenAI.
Rumours suggest that OpenAI is getting closer to SoftBank and is working on the "Project Stargate" with them. There are also discussions about OpenAI giving up Microsoft's server capacities. This strategic shift towards independence from OpenAI is a significant development in the AI landscape.
Microsoft sees voice as "the future interface for AI companions." As the race for AI dominance intensifies, it will be interesting to see how these developments unfold in the coming years. Both Microsoft and OpenAI have big ambitions for the future and aim for further progress in the field.
