De-identification

noun

De-identification is the process of modifying or removing personal data from a dataset so that individuals cannot be identified. This is often done to comply with privacy laws like the GDPR or to allow data sharing without exposing personal information.

It typically involves techniques like anonymization (removing direct identifiers like names or email addresses) or pseudonymization (replacing them with codes). For example, a healthcare dataset might replace patient names with random IDs to allow research while protecting privacy.

For businesses, de-identification can reduce legal risks by making data less vulnerable to breaches. However, it must be done carefully—if too much detail is removed, the data may lose its value. A 2023 study found that 60% of anonymized datasets could still be re-identified with enough effort, so strong methods are essential.

A marketing firm de-identifies customer purchase data before selling it to researchers, keeping transaction details but removing names and addresses.

See also: anonymization · Pseudonymization · GDPR

Découvrez nos autres services

ElpisIA, c’est tout un univers — explorez nos autres services.