AI Ethics

Introduction

• AI Ethics is a branch of applied philosophy. It deals with the moral principles and values that guide the development and deployment of AI. It answers the question: "Just because we can build this AI, should we? And if so, according to what values?"

Core Principles

Privacy

• To ensure each individuals can control over personal information and how it is used

• For data collection, it is required user approval for the collection of data

• During data preprocessing, some sensitive data may need to be filtered or masked

Transparency

• To resolve "Black Box" problem. Users should understand how an AI arrived at a decision, especially in high-stakes areas like healthcare or criminal justice.

• How the data is collected and used. How the model be trained / fine-tuned and answered

• To draft the benchmark and scope, which question can be answered, what is the confidence score and limitation

Accountability

• AI accountability is not just about assigning blame; it is about taking responsibility for the entire lifecycle of the system

• Organizations must explicitly define who owns the risk before the AI is deployed.

  • The Developers: Responsible for testing the model for bugs and bias.

  • The Product Owners: Accountable for the decision to deploy the model in the real world.

• Does the process follow the legal compliance?

• How government law secure the safety of using AI?

Fairness

• AI fairness is to treat everyone equally, avoiding discrimination and bias

• During data collection, is the data variety high enough?

• Is the model be fine-tuned by updated data?

• Does the model run the fairness check across different scenarios?

AI Governance

• Artificial intelligence (AI) governance refers to the processes, standards and guardrails that help ensure AI systems and tools are safe and ethical.

• AI governance is essential for reaching a state of compliance, trust and efficiency in developing and applying AI technologies.

• Effective AI governance includes oversight mechanisms that address risks such as bias, privacy infringement and misuse while fostering innovation and building trust.