Privacy-enhancing technologies,
where they fit.
PET is a family of techniques, not one product. Here is what mintBlue uses, mapped to the OECD's four groups, and where another provider is the better choice.
Data obfuscation
Each party opens only what it is entitled to.
Encrypted data processing
Learn whether you hold the same case, without handing over source data.
Federated and distributed analytics
Source records stay with their holder.
Data accountability
Proof of what was shared, by whom and on which mandate.
PET limits what is exposed. It does not decide what may be shared.
The OECD groups privacy-enhancing technologies in four families: data obfuscation, encrypted data processing, federated and distributed analytics, and data accountability tools. They let organisations collect, process, analyse and share data while keeping it confidential.
Who may see what, and why, is agreed first. PET is not a silver bullet.
What PET does not do
- Create a legal basis to share
- Decide the purpose of an exchange
- Decide who may see what
- Replace the agreement between the parties
What PET does
- Limits what each party sees
- Keeps data confidential while it is checked or matched
- Leaves proof of what was shared, by whom and on which mandate
The OECD's four groups, and where mintBlue fits
Data obfuscation
- Where another provider fits better:
- Synthetic data and differential privacy for research and model training. We do not offer these.
- What mintBlue offers:
- Selective disclosure: a key per field, so each party opens only what it is entitled to, often a yes or a no, while the source record stays with its holder.
Encrypted data processing
- Where another provider fits better:
- General analytics on combined encrypted data (MPC or homomorphic encryption platforms). That is their core; we stream to it.
- What mintBlue offers:
- Matching on blinded tokens, so organisations can learn whether they hold the same case without handing over source data. Our event streams can feed an MPC platform, so it computes over new events instead of whole datasets.
Federated and distributed analytics
- Where another provider fits better:
- Federated learning platforms.
- What mintBlue offers:
- Source records stay in each holder's system, and our event streams can feed distributed analytics, but we do not run federated analytics or train models across holders.
Data accountability
- Where another provider fits better:
- This is mintBlue's primary focus.
- What mintBlue offers:
- The core of mintBlue: signed events, a tamper-evident audit trail, mandates checked at every action, and audit-only access so an auditor checks what happened without opening the data.
Sources: OECD (2023), Emerging privacy-enhancing technologies; UK ICO, Privacy-enhancing technologies guidance; ENISA, Data protection engineering.
Frequently asked questions
PET
Next step
Start with one exchange.
Show us one exchange and what your auditor needs to see. We scope a pilot around it, with a go or no-go after each step.
Next step
Start with one exchange.
Show us one exchange and what your auditor needs to see. We scope a pilot around it, with a go or no-go after each step.