Microsoft Puts Human Control First with New 'Humanist AI' Code

Microsoft has published a new 'Humanist AI' code of conduct that emphasizes human control over artificial intelligence. The company's CEO, Satya Nadella, said he supports calls to pace AI development and echoed Microsoft's motto: 'People matter more than AI.'

The code rejects the idea of a 'race to produce an all-purpose superintelligence,' which could potentially lead to human loss of control. Instead, it advocates for designing AI as a tool that increases human agency rather than fostering dependence.

According to Microsoft's Chief Responsible Tech Officer, Suleyman, this approach means being willing to compromise on capabilities and speed in order to preserve safety and control. He believes that humanist superintelligence is possible and can be achieved by prioritizing human decision-making roles.

The company plans to revise the code following a six-week public consultation and will use the resulting version to guide model development beginning in 2027.