Synthetic Data: Opportunities and Risks for Financial, Insurance, and Supervisory Innovation

Synthetic data is becoming increasingly important for financial and insurance innovators, as well as supervisors.

A recent World Economic Forum report prompted me to reflect further on its role in innovation and financial services.

What is Synthetic Data?

The WEF defines synthetic data as artificially generated data that mimics the statistical properties, structure, and distribution of real-world data.

It can:

• Fill data gaps
• Protect privacy
• Enable the testing of new scenarios

Synthetic data offers a scalable and cost-effective alternative when real data is limited, sensitive, or biased.

Applications in Finance and Insurance

For financial and insurance markets, synthetic data has wide-ranging potential:

• Correcting gaps and bias: addressing scarcity or demographic imbalances, improving fairness in lending, healthcare or insurance.
• Financial inclusion: reducing bias in credit/insurance decisioning.
• Crisis modelling: enabling simulation of blackouts, pandemics, or other systemic shocks.
• Safe experimentation: providing cost-effective datasets for training, testing, and developing models without exposing sensitive customer data.

Risks and Challenges

Synthetic data is not without risks. If poorly generated or governed, it can:

• Reinforce existing biases from the original datasets
• Mislead decision-makers with inaccurate results
• Leak sensitive information if improperly anonymized

To mitigate these risks, accuracy, traceability, and clear labelling are essential. Without robust governance, synthetic data could undermine trust and model performance.

Supervisory Use Cases

Synthetic data is also valuable for supervisors.

During my supervisory work, for example, we developed the EU Digital Finance Platform and Data Hub. The hub provides participating firms with access to synthetic supervisory data to test new solutions and train AI/ML models.

Similarly, the UK Financial Conduct Authority recently published a report on governance considerations for synthetic data. It provides practical insights for supervisors and innovators alike.

Data Protection Considerations

Data protection remains crucial. The European Data Protection Supervisor (EDPS) has recently highlighted both the positive and negative implications of synthetic data. Ensuring privacy while maintaining utility is a delicate balance.

Conclusion

Synthetic data offers significant opportunities for innovation in financial and insurance services supporting inclusion, enabling safe experimentation, and improving resilience.

At the same time, it raises pressing questions around governance, fairness, and data protection.

For innovators and supervisors alike, careful design, transparent labelling, and robust oversight are essential to ensure synthetic data becomes a trusted enabler rather than a new source of risk.

📬 Subscribe to my newsletter here
📌See how I can help you here
🔗 Follow me on LinkedIn
📧 andres@insurtech4good.com