Privacy Meets AI

Inspiration

I wanted to explore how AI can help protect user privacy, especially with sensitive information like emails, phone numbers, NRICs, and addresses. With generative AI becoming mainstream, safeguarding personal data is crucial.

What it does

Privacy Meets AI detects Personally Identifiable Information (PII) in text and sanitises it before sending it to AI services. It highlights PII types, calculates a risk level, and outputs a safe-to-use version of the text.

How we built it

I built this tool using Python, using libraries like re for regex pattern matching, date-time for timestamps, and optionally Flask for a web UI. The detection uses Singapore-specific PII patterns such as phone numbers, NRICs, addresses, emails, and postal codes.

Challenges we ran into

Creating accurate regex patterns for Singapore-specific PII and designing optional UI features for a better user experience were some of the main challenges.

Accomplishments that we're proud of

I successfully built a functional PII detection and sanitisation tool tailored for Singapore, and designed the code to be easily expandable for other PII types or regions.

What we learned

Regex can be powerful but has limitations for complex patterns. I learned how to structure code for scalability and user-friendly interactions, and the importance of testing with local data and edge cases.

What's next for Privacy Meets AI

Develop a web-based UI using Flask, expand support for other regions beyond Singapore and add AI-based detection for PII patterns that regex cannot catch.

Built With

• anaconda
• jupyter
• python