Microsoft CEO Satya Nadella issued an internal memo on September 14, emphasizing that AI development must remain under human control. He advocated for broader third-party testing and warned that companies must prioritize safety to avoid losing their operational licenses as risks increase.
Nadella outlined three strategic pillars for Microsoft. First, he stressed the need for a balanced ecosystem supporting both open and closed-source models. Second, he highlighted the importance of customer data sovereignty, noting that the Foundry platform enables enterprises to maintain control over their knowledge without relying on a single provider.
The third pillar focuses on community impact, with Nadella citing the Quincy data center as a model for local infrastructure development. Additionally, Microsoft recently released a 37-page draft of its human-centric AI model code of conduct, as the company seeks to align its practical actions with its stated safety goals.
The memo follows recent industry discussions regarding AI safety. Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have both expressed support for external monitoring and potential slowdowns in deployment to ensure that AI systems remain aligned with human interests and safety standards.
