Close-up of a scientist holding a tray of petri dishes with purple cultures, wearing gloves and safety gear.

Privacy Enhancing Technologies (PETs) framework

An assessment tool that recommends the right privacy technologies based on your specific data collaboration requirements.

Start Assessment

Book a Call

Privacy Technology Assessment
Answer all questions to get personalized recommendations
REGULATORY REQUIREMENTS
Which privacy-regulation landscape best describes your organisation’s compliance obligations?
Which privacy-regulation landscape best describes your organisation’s compliance obligations?
COMPUTATION TYPE
During collaborations with external partners, what form of joint data processing do you most often perform?
During collaborations with external partners, what form of joint data processing do you most often perform?
COLLABORATORS
How many distinct parties participate in the collaboration?
Collaborators include any party providing data inputs or providing models, queries & analytics
How many distinct parties participate in the collaboration?
LATENCY
Does the system need to produce responses in real time?
Does the system require batch processing or real-time processing
Does the system need to produce responses in real time?

What are Privacy Enhancing Technologies?

Privacy Enhancing Technologies (PETs) are a collection of technologies that enable organizations to process, analyze, and share data while preserving privacy and confidentiality.

Multi-Party Computation (MPC)

Enables multiple parties to jointly compute functions over their inputs while keeping those inputs private.

Multi-Party Computation (MPC)

Enables multiple parties to jointly compute functions over their inputs while keeping those inputs private.

Fully Homomorphic Encryption (FHE)

Allows computations to be performed on encrypted data without needing to decrypt it first.

Fully Homomorphic Encryption (FHE)

Allows computations to be performed on encrypted data without needing to decrypt it first.

Zero Knowledge Proofs (ZKP)

Proves the truth of a statement without revealing any information beyond the validity of the statement itself.

Zero Knowledge Proofs (ZKP)

Proves the truth of a statement without revealing any information beyond the validity of the statement itself.

Trusted Execution Environment (TEE)

Provides secure areas of a processor to guarantee code and data are protected with respect to confidentiality.

Trusted Execution Environment (TEE)

Provides secure areas of a processor to guarantee code and data are protected with respect to confidentiality.

Synthetic Data

Artificially generated data that maintains statistical properties of real data without exposing sensitive information.

Synthetic Data

Artificially generated data that maintains statistical properties of real data without exposing sensitive information.

Data Anonymisation

Process of removing or modifying personally identifiable information to prevent identification of individuals.

Data Anonymisation

Process of removing or modifying personally identifiable information to prevent identification of individuals.

Differential Privacy

Mathematical framework for measuring and limiting privacy loss when statistical queries are performed on datasets.

Differential Privacy

Mathematical framework for measuring and limiting privacy loss when statistical queries are performed on datasets.

Federated Learning

Machine learning technique that trains algorithms across decentralized devices without exchanging raw data.

Federated Learning

Machine learning technique that trains algorithms across decentralized devices without exchanging raw data.

Case Studies

-2.0% to 3.0%

minimal error deviation ensuring high utility

Digital Advertising In A Paradigm Without 3rd Party Cookies

Meta implemented differential privacy for targeted advertising while protecting user privacy, enabling campaign measurement without compromising individual user data.

DP

≈1.4%

re-identification risk with 5% distance to real data

Generating Synthetic Data for Analysis and Research

Kajima Corporation developed synthetic data generation for construction and real estate analysis, enabling research while protecting sensitive project and customer information.

Synthetic

>90%

relative increase to baseline conversion rates

Enhancing Customer Engagement With Privacy Preserving AI

Ant International implemented federated learning for collaborative model training across business units while keeping customer data decentralized and private.

FL

2 decimals

precision of computations matched the accuracy of non-encrypted computations

Privacy Preserving Attribution and Measurement

TikTok developed PrivacyGo, an open-source solution using secure multi-party computation for privacy-preserving advertising attribution and measurement without compromising individual user privacy.

Impressive metrics

MPC

90%

lift in performance due to improved targeting

Collaboration on First Party Data to Enable Customer Activation

SPH Media implemented trusted execution environments for secure collaboration on first-party customer data while maintaining confidentiality and integrity.

TEE

Zero

loss of accuracy vs clear text

Pseudonymising employee data for recruitment analytics

Case study on pseudonymising employee data for recruitment analytics while protecting employee privacy through proper anonymisation techniques.

Anonymised Data

Preventing Financial Fraud Across Different Jurisdictions With Secure Data Collaborations

Mastercard used fully homomorphic encryption for secure sharing of financial crime intelligence across institutions, enabling collaborative fraud detection while keeping transaction data encrypted.

FHE

BIS Project Mandala: Streamlining cross border transaction compliance

Compliance-by-design approach to streamline cross-border compliance processes for financial institutions and explores real-time policy and regulatory compliance monitoring for central banks & other regulators.

ZKP

MPC