Democratizing enterprise synthetic data privacy protection

With enterprise-grade privacy features and provable anonymity, synthetic data adoption is a breeze.

synthetic data privacy

How do you ensure the privacy
and security of synthetic data?

The popularity of synthetic data is on the rise. For many teams, synthetic data represents a safe way to innovate while complying with data privacy regulations. But the use of synthetic data does not exempt companies from assessing re-identification risks, as required by the GDPR.

Our privacy evaluators enable data teams to assess the robustness of synthetic data against re-identification attacks in a single click. With Statice Evaluators, enterprises can simplify the adoption of synthetic data, easily comply with anonymization requirements, and protect individuals' privacy.

The privacy of synthetic data

The synthetic data generation process irreversibly breaks one-to-one links between synthetic and real data records. This irreversible approach reduces the re-identification risk.

synthetic data privacy process

However, the deep learning models used for synthetic data generation might memorize features during the synthesization process. Ultimately, these memorized patterns can be reproduced in the synthetic data, leading to privacy leaks.

The risk assessment of synthetic data is left up to each company's discretion. Due to the limited recommendations available, implementing a risk assessment becomes a challenge, and puts individuals' privacy at risk. To address this gap, we developed a set of evaluations so you can measure the re-identification risks of synthetic data.

The Statice Privacy Evaluators

This evaluation kit provides a data-driven assessment of the protection against all known re-identification risks related to synthetic data. 

re-identification risk

Benefits for data stakeholders

Fast deployment of synthetic data for data science teams thanks to a ready-to-use privacy kit

Easy documentation of the robustness of data protection to ease compliance processes for DPOs

State-of-the-art protection of data subjects’ privacy and business secrets

Linkability evaluator

This Linkability evaluator measures the re-identification risk by evaluating how much help the synthetic data gives to an attacker who wants to establish links between records belonging to the same individual. 

data privacy impact assessment synthetic data
data privacy impact assessment synthetic data

Inference evaluator

The Inference evaluator detects privacy leaks by assessing how much information an attacker with partial knowledge of the original data could gain by seeing the synthetic data. If the synthetic data provides more information about the training data than it provides about held-out test data, there is a privacy leak. 

Singling Out evaluator

The Singling Out evaluator analyzes the probability of isolating records that would identify an individual. It measures the robustness of the dataset against a scenario in which an attacker would use the synthetic data to determine the presence of an individual in the dataset.

data privacy impact assessment synthetic data

European enterprises reduce time to value with the Statice Evaluators

Provinzial's data team had to provide a high privacy-preserving solution that satisfied not only the GDPR but also internal privacy requirements. Other tools posed a greater risk of re-identification and offered no means of assessing privacy. With the Statice evaluators, Provinzial shortened the process of evaluating data privacy risks by three months.

⟶ Read more

Got a question
for our security and privacy team?

Get in touch