Preserving privacy on data collaboration with AWS Clean Rooms

Presented at AWS re:Inforce 2024

Transcript

1. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved.
2. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Preserving privacy on data collaboration with AWS Clean Rooms Richard Fan C O M 2 2 1 AWS Security Hero
3. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Data collaboration Data subjects Data custodians Data collaborators Data outputs Risk of exposure Amount of information
4. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Challenge 1: PII Personal identifiable information (PII) – For example, SSN, passport number, credit card number • Well-defined • Easy to identify, remove Amazon Macie
5. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Challenge 2: Usability tradeoff For example, joint member promotion between airline and hotel Flights taken Passport no. Passport no. Nights stayed 10 123456789 123456789 1 1 987654321 987654321 20 8 112233445 112233445 15 Airline Hotel
6. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Flights taken Passport no. Passport no. Nights stayed 10 123456789 123456789 1 1 987654321 987654321 20 8 112233445 112233445 15 Airline Hotel Challenge 2: Usability tradeoff For example, joint member promotion between airline and hotel PII Impossible to collaborate
7. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. AWS Clean Rooms Data collaborators Data outputs Data collaborators AWS Clean Rooms Analysis rules
8. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. AWS Clean Rooms analysis rules Aggregation *** *** *** *** Blind match Cryptographic Computing for Clean Rooms (C3R)
9. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Challenge 3: Output privacy Data subjects Data collaborators Data outputs Amount of information Auxiliary data Re-identification
10. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Differential privacy Privacy budget Noise
11. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. A look into AWS Clean Rooms
12. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. A look into AWS Clean Rooms
13. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. A look into AWS Clean Rooms
14. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. A look into AWS Clean Rooms
15. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Takeaways • Understand your data throughout the pipeline • Usability vs. privacy • Output privacy is also important
16. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Learn more AWS Clean Rooms hands-on lab Blog post series on differential privacy AWS Clean Rooms documentation
17. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Thank you! © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Thank you! Please complete the session survey in the mobile app Richard Fan @richardfan1126 richardfan1126