Microsoft's Responsible AI principles of Accountability and Reliability & Safety emphasize human oversight for AI systems. For high-impact scenarios like customer communications, it is crucial that a human reviews and takes responsibility for the final content. This practice ensures accuracy, appropriateness, and maintains brand voice. Communicating the AI's limitations is part of the Transparency principle, ensuring users understand that Copilot is an assistant, not an infallible decision-maker. This approach mitigates risks associated with unverified AI-generated content being sent to customers.

B. This is an operational tactic, not a responsible AI principle. It doesn't address the quality or appropriateness of the content itself.

C. Disabling logging contradicts the principles of Transparency and Accountability, as it prevents auditing or understanding how the AI is being used.

D. This practice violates the principle of Accountability by removing human oversight and judgment, which is essential for responsible AI use.

1. Microsoft. (2023). Microsoft Responsible AI Standard

v2. Section 1.6

"Accountability

" Principle 6

Guideline 1: "Establish meaningful human oversight."

2. Microsoft Learn. (2024). Get started with Copilot for Microsoft 365. Section: "Our commitment to responsible AI." Paragraph 2: "We believe that when you create AI

you must do it responsibly. That means putting people first and building in safeguards to protect them."

3. Microsoft Learn. (2024). Data

Privacy

and Security for Microsoft Copilot for Microsoft 365. Section: "Microsoft's commitment to responsible AI." Paragraph 3: "The Accountability principle requires that we hold ourselves accountable for operating our AI systems in a way that is consistent with our AI principles."