Satya Nadella announced that Microsoft will publish a "Code of Conduct" for its MAI models and opened, on September 14, a six-week public comment period on the draft. The document explicitly bans its models from carrying out cyberattacks, assisting with nuclear weapons development, or generating deepfakes.
Nadella summed up the logic of the code in one line: "any pursuit of superintelligence has to be grounded in the core principle that if the AI we build doesn't help humanity and isn't under human control, it's not worth pursuing." It's the first time Microsoft has formalized in writing — as a public governance document rather than a PR statement — the operational limits of its own first-party models (MAI), the family that competes directly with the OpenAI, Anthropic, and Google models Microsoft also resells via Azure.
The Code of Conduct draft bars MAI models from carrying out cyberattacks, assisting with nuclear weapons development, or generating deepfakes. On the governance side, Nadella argued that the rules governing AI "shouldn't be controlled by a handful of entities" and instead called for those rules to be shaped by a broad mix of countries, companies, and academics — a stance that contrasts with the more closed, unilateral approach other labs have taken in setting their own usage policies.
For companies in Guatemala and Latin America using Copilot, Azure AI Foundry, or any Microsoft product built on MAI models, this code of conduct — once finalized after the public comment period — will likely translate into more explicit, documented usage restrictions in the terms of service, which actually makes internal compliance work easier: it's simpler to audit and justify using a model whose behavioral limits are published and subject to outside review than one whose safety policies are opaque. Companies already evaluating which model provider to use for regulated use cases (healthcare, finance, government) should start asking their vendors for equivalent governance documents as a selection criterion, not just price and technical capability.
Carlos Montiel is an enterprise AI solutions architect. He implements LLMs, Agents, RAG, and orchestrators for companies across Guatemala and Latin America. Reach out for a consultation.
Contact Carlos Montiel