AI — Ethics

AI Ethics Principles

Principle	Description
Fairness	Avoid bias
Transparency	Explainable AI
Privacy	Data protection
Accountability	Responsibility
Safety	Avoid harm

Bias in AI

• Training data bias
• Algorithmic bias
• Deployment bias

Fairness Techniques

from fairlearn.metrics import MetricFrame
from sklearn.metrics import accuracy_score

metric_frame = MetricFrame(
    metrics=accuracy_score,
    y_true=y_true,
    y_pred=y_pred,
    sensitive_features=protected_attribute
)

Explainable AI

import shap

explainer = shap.Explainer(model)
shap_values = explainer(X_test)
shap.summary_plot(shap_values, X_test)

Mini Practice

1. Understand AI ethics
2. Check for bias
3. Implement fairness metrics
4. Use explainable AI

Up Next

Continue with Bias — bias in AI.