Personal data anonymization with model verification

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 31st, 2026

Detected March 31st, 2026

Summary

USPTO granted Patent US12591708B2 to Twilio Inc. on March 31, 2026. The patent covers methods and systems for anonymizing personal data and verifying the effectiveness of anonymization techniques by comparing results against random guessing thresholds. The patent includes 17 claims and names five inventors.

View original document View source feed page

What changed

USPTO issued Patent US12591708B2 to Twilio Inc. for a method of anonymizing personal data and verifying the anonymization by comparing results to a random guessing baseline. If the difference falls below a predetermined threshold, the anonymization is deemed safe for use with machine-learning models. The patent claims techniques to prevent re-identification of individuals from anonymized datasets.

Technology companies and data handlers should note this patent establishes intellectual property boundaries in the data anonymization and privacy-preserving machine learning space. The filing date was November 11, 2020 (Application No. 17095185). This is an informational patent grant with no compliance obligations; it primarily affects companies developing similar anonymization technologies who should review potential patent clearance implications.

Source document (simplified)

← USPTO Patent Grants

Personal data anonymization with model verification

Grant US12591708B2 Kind: B2 Mar 31, 2026

Assignee

Twilio Inc.

Inventors

Aaron Beach, Liyuan Zhang, Tiffany Callahan, Mengda Liu, Shruti Vempati

Abstract

Anonymization is the process to remove personal information from the data. Once the data is anonymized, the data may be used for creating machine-learning models without the risk of invading anyone's privacy. One of the keys to data anonymization is to make sure that the data is really anonymized so nobody could use the anonymized data to obtain private information. Different techniques for data anonymization are presented. Further, the data anonymization techniques are tested for true anonymization by comparing the results from these techniques to a random method of guessing. If the difference in the results is below a predetermined threshold margin, the data anonymization techniques are safe and ready for use.