Which principle emphasizes that AI systems should be understandable and provide clear information on how they work?

The principle of Transparency directly addresses the need for AI systems to be understandable. It mandates that users and stakeholders are provided with clear, accessible information about how an AI system functions, including its capabilities, limitations, and the data it uses. For a tool like GitHub Copilot, this means being clear about how it generates code suggestions and what its potential pitfalls are. This principle is fundamental to building trust and enabling users to make informed decisions when interacting with the AI.

A. Fairness: This principle focuses on ensuring AI systems treat all user groups equitably and avoid perpetuating societal biases, not on explaining how the system works.

C. Inclusiveness: This principle is about designing AI systems that are accessible and beneficial to a diverse range of people, which is distinct from operational clarity.

D. Accountability: This principle concerns establishing who is responsible for the AI's outcomes and actions, rather than making its internal processes understandable.

1. Microsoft Responsible AI Standard, v2. (Official Vendor Documentation). Part 1: The Microsoft Responsible AI Standard, Section: "The six Microsoft responsible AI principles," p. 6. The document states, "The principle of transparency is about providing context and information that helps people understand AI systems’ behavior."

2. GitHub Copilot Trust Center. (Official Vendor Documentation). Section: "How does GitHub Copilot work?". This section exemplifies transparency by explaining the underlying models (e.g., OpenAI Codex), the data used for training, and the process of generating suggestions, which aligns with the principle's goal of making the system understandable.

3. Stanford University, Human-Centered Artificial Intelligence (HAI). (University Courseware/Publication). "The AI Index 2023 – Chapter 5: Policy and Governance." This report discusses global AI policy trends, where transparency is consistently cited as a core principle for AI regulation and governance, emphasizing the need for systems to be explainable and their operations clear to auditors and users.