Glossary
Differential privacy
Quick answer
A mathematical technique that adds carefully measured noise to data or results so individuals cannot be singled out.
It gives a provable privacy guarantee, controlled by a setting often called epsilon. The U.S. Census Bureau and large tech companies use it for statistics.
It works well for aggregate statistics but can reduce the usefulness of detailed records, so most business data licensing relies on redaction and de-identification instead.
Why Differential privacy matters when licensing company data to AI
Differential privacy is a mathematical technique that adds carefully measured noise to data or results so individuals cannot be singled out. It gives a provable guarantee, controlled by a setting often called epsilon. For licensing, it suits aggregate statistics, but it can reduce the usefulness of detailed records, so most business data licenses rely on redaction and de-identification instead.
- 01Your work records
- 02Personal details removed
- 03You approve the release
- 04Licensed to AI developers
Related terms
Related resources
- IndustryBPO & contact centers data
- QuestionCan anonymized data be licensed?
- QuestionCan CRM data be licensed?
- InsightHow do I de-identify sales call recordings for AI training?
- InsightHow do I de-identify video recordings for AI training?
- InsightHow do I de-identify voice and audio data for AI training?
- SolutionData partnerships between businesses and AI developers
- IndustryRecruiting & staffing data
- DataSales and CRM data
- DataCall recordings
- GlossaryPersonally identifiable information
- GlossaryData processing agreement
See if your company qualifies
A short company assessment. No data uploads are needed.