Ethics & Privacy

Data scientists work with people's data. Follow the principles: consent, minimisation, purpose limitation, security.

Audit models for BIAS across gender, race, age and other sensitive groups.

Example 1 (python)

# Slice metrics by demographic group
# to check for disparate impact

Fairness starts with per-group metrics.

Example 2 (python)

# Anonymise/aggregate PII; consider differential privacy

Modern privacy tooling.

Key points

💡 Note: A '99% accurate' model that discriminates against a demographic is not a success — it's a legal and ethical failure.