What is synthetic data, and how is it fundamentally different from masking or anonymizing real data?

Synthetic data is entirely artificial records of fictitious individuals generated from a statistical model — not real data that's been modified.

Synthetic data generation creates completely artificial datasets that can replace real, sensitive data. Rather than masking or adjusting existing records (the traditional anonymization approach), it generates entirely new datasets populated with fictitious individuals.

The core principle: replace sensitive values with artificially generated alternatives derived from statistical models. It mimics sampling from a population — but not with actual people, with realistic synthetic individuals that statistically resemble the original population.

Tip: Anonymization edits the real people's records; synthetic data invents new "people" who never existed but behave like the real ones in aggregate.

Go deeper:

• Synthetic data (Wikipedia) — generation methods, distinguishing modeled artificial records from edited real ones.