Patricia Thaine

Organizations often choose PII detection vendors without a fair way to measure accuracy, leading to costly compliance and security risks. This post outlines Private AI’s research-based benchmarking methodology—covering datasets, precision/recall metrics, and real-world testing—so you can make informed decisions about the tools that protect your sensitive data.

Traditional methods like regex fall short when it comes to detecting hidden PII in real-world contexts such as call transcripts and unstructured data. This post explores the risks of missed PII, why open-source solutions struggle, and how AI-driven approaches provide the accuracy and scalability needed to protect sensitive information and maintain trust.

I’m thrilled to announce Private AI 4.0. This release redefines how healthcare businesses harness their unstructured data while maintaining fine-grained privacy controls and the highest standards of compliance. With Private AI 4.0, we are turning challenges into opportunities by enabling analytics confidently and securely to improve patient outcomes, accelerate research, and drive operational efficiency.

In the fast-evolving world of AI, where data is at the core of decision-making, one challenge remains constant: ensuring appropriate privacy and protection of an individual’s and an organization’s data. At Private AI, we’ve been working on solving this challenge with the introduction of Coreference Resolution as part of our 4.0alpha release.

Discussing developer responsibility, Bill C-11, positive consent, and the importance of Privacy by Design

Understanding key tech for data protection regulation compliance

We cover the basics of homomorphic encryption, followed by a brief overview of open source HE libraries and a tutorial on how to use one of those libraries (namely, PALISADE).

A number of people ask us why we should bother creating NLP tools that preserve privacy. Apparently not everyone spends hours thinking about data breaches and privacy infringements.

A very brief overview of privacy-preserving technologies follows for anyone who’s interested in starting out in this area. I cover symmetric encryption, asymmetric encryption, homomorphic encryption, differential privacy, and secure multi-party computation.

General-purpose cloud tools miss a large share of sensitive data, leaving organizations exposed to compliance and security risks. In this post, we share benchmark results comparing Private AI’s purpose-built PII detection to AWS, Azure, Google DLP, and Microsoft Presidio—highlighting why specialization delivers far higher recall and greater protection against data breaches.