AAISM: Advanced in AI Security Management

De-identification reduces privacy risk but does not permanently eliminate the possibility of re-identification. A monitoring program with privacy audits and adversarial testing evaluates whether evolving AI capabilities can re-identify individuals and provides evidence for risk treatment. NIST recommends evaluating the risks created by de-identification and performing re-identification studies to gauge that risk.