On Thursday, Microsoft announced the creation of a new advanced artificial intelligence research team, called the MAI Superintelligence Team, led by Mustafa Suleyman, who is currently in charge of the Microsoft AI division. The stated goal is to develop "practical and controllable" superintelligence geared toward concrete applications in the fields of education, healthcare, and renewable energy. The project stands out for its desire to produce AI rooted in reality, with targeted uses such as digital learning companions, medical diagnostic tools, and energy optimization solutions.
This announcement comes amid increased competition between tech giants, just a few months after the creation of Meta Superintelligence Labs. The concept of superintelligence, generally associated with artificial cognitive abilities that exceed those of humans, is approached with caution by Suleyman, who insists on a measured and responsible approach. A former co-founder of DeepMind who worked at Inflection AI before joining Microsoft, he embodies the group's desire to gradually break free from its dependence on OpenAI, while capitalizing on recently acquired talent.
Microsoft, which powers Bing and Copilot with OpenAI models, also hosts its operations on Azure, but is now exploring internal and external alternatives, such as models from Alphabet or Anthropic. While some analysts question the profitability of massive investments in AI, Microsoft is demonstrating a strategy focused on measurable societal impact with this initiative. The company thus intends to assert its position as a technology leader, while promoting a structured and useful vision of superintelligence.
