● Covered by 1 source · 1 reportMedium impact1 positive
Rampart Open-Sourced for On-Device PII Redaction in Browsers
Key points
- Rampart is an open-source, on-device PII filtering system.
- It uses regular expressions and MiniLM for redaction.
- All processing occurs in the browser before data is sent.
- Aims to prevent PII from leaving the user's device.
Introduction of Rampart
Rampart, a new on-device personal information filtering system, has been open-sourced. This tool is designed to act as a first line of defense, ensuring that personal identifiable information (PII) does not leave a user's device, particularly when interacting with online services like chatbots. The core principle behind Rampart is that private information should remain on the device.
Addressing PII Privacy Challenges
The development of Rampart addresses challenges associated with PII removal, such as the difficulty in verifying privacy guarantees from remote AI servers and the large size of existing PII removal models. Remote servers pose risks like accidental logging of sensitive data, zero-day vulnerabilities, and insider threats. Additionally, models like OpenAI Privacy Filter are approximately 2.8GB, making them impractical for browser-based, on-device use due to long download times.
How Rampart Functions
Rampart operates entirely within the browser, processing messages between typing and sending, without involving a remote server. It employs a two-stage filtering process. The first stage uses deterministic regular expressions to identify and redact structured information such as Social Security numbers, credit card numbers, phone numbers, and email addresses. The second stage utilizes MiniLM, a small language model, to understand context and redact less structured PII like names and street addresses.
Impact on Data Privacy
By performing PII redaction locally, Rampart offers a method for users to maintain control over their sensitive data. This approach aims to mitigate the risks associated with transmitting personal information to external servers, providing a more verifiable privacy solution for online interactions.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Primary sources
npm @nationaldesignstudio/rampartReporting from
Rampart, an on-device personal information filtering system, has been open-sourced to prevent sensitive data from leaving a user's device when interacting with chatbots. It combines regular expressions and a small language model (MiniLM) to redact PII locally, addressing privacy concerns with remote servers and large PII removal models.