TECHNOLOGY

Keeping Humans in Charge: Microsoft's Plan for Friendly Superintelligence

Thu Nov 06 2025

Microsoft's AI division is making big promises about the future of superintelligent AI. They aim to create AI that is not just smart but also safe and helpful to humans.

A Helpful Assistant, Not a Boss

This new AI would act like a helpful assistant, not a boss. It would have limits and be designed to help people in various areas such as:

  • Learning
  • Healthcare
  • Science

Controlled and Safe

The head of Microsoft AI, Mustafa Suleyman, assures that this AI won't be a wild, uncontrolled force. Instead, it will be a tool that people can use and control. Microsoft is not in a rush to create this superintelligent AI.

The OpenAI Deal

However, there's a catch. Microsoft has a deal with OpenAI, allowing them to use OpenAI's technology to build their own superintelligent AI.

Big Plans for the Future

Suleyman envisions this AI as a friend that helps people:

  • Learn
  • Be more productive
  • Assist doctors and scientists in making new discoveries

The Challenge of Control

Despite these plans, some may wonder if Microsoft can keep this AI under control. After all, AI is getting smarter every day. Ensuring it remains friendly and helpful is a big challenge.

questions

    How will Microsoft AI address concerns about the competition with OpenAI in the pursuit of AGI?
    What are the potential risks associated with Microsoft AI's vision of 'humanist superintelligence'?
    If the AI is designed to be a 'subordinate, controllable AI,' will it still be able to win at video games against its human counterparts?

actions