Synthetic data is artificially generated data designed to resemble or supplement real-world data.
It can help with privacy, testing, data scarcity, and class imbalance, but it must be evaluated for realism, bias, leakage, and fitness for purpose.