Microsoft Enters AI Development Race

Microsoft is stepping up its AI game by developing a powerful in-house language model aimed at competing with those from Alphabet’s Google and OpenAI, as reports from The Information on Monday. This strategic move is part of Microsoft’s broader plan to strengthen its AI capabilities and maintain its competitive edge in the industry.

Dubbed internally as MAI-1, the new AI model is under the guidance of Mustafa Suleyman, a recent addition to Microsoft’s team. Suleyman, known for his co-founding role at Google DeepMind and previous CEO position at AI startup Inflection, is leading this ambitious endeavor, according to two Microsoft insiders familiar with the project.

The specific function of the model remains undecided and will depend on its performance. As the report outlines, Microsoft may showcase the new model as early as its Build developer conference later this month.

Microsoft opted not to provide a statement when approached for comment.

According to a report, MAI-1 will surpass Microsoft’s previous smaller, open-source models in size, making it more costly to develop and maintain.

Last month, Microsoft introduced a smaller artificial intelligence model named Phi-3-mini, aiming to appeal to a broader range of clients with cost-effective options.

Having invested billions in OpenAI, Microsoft has integrated the ChatGPT maker’s technology across its productivity software, granting it an early advantage in the generative AI competition.

According to the report, Microsoft has allocated significant clusters of servers equipped with Nvidia’s graphic processing units and extensive datasets to enhance the model.

According to the report, MAI-1 is expected to boast approximately 500 billion parameters, a significant leap from Microsoft’s Phi-3 mini’s 3.8 billion parameters. In comparison, OpenAI’s GPT-4 reportedly reaches one trillion parameters.

In March, Microsoft appointed Suleyman to lead its newly established consumer AI unit and hired several former employees of Inflection.

Although the new model isn’t a direct continuation of Inflection’s work, it may leverage training data accumulated by the startup, the report noted.

Read More: Microsoft Bans US Police Departments From Using Their AI Facial Recognition Tool