Policy & Ethicschatbotsenterprise modelstraining leakageprivacy
Google Engineer Advises AI Privacy Best Practices
7.0
Harsh Varshney, a 31-year-old Google engineer, outlines four habits for protecting personal and company data when using AI chatbots in an as-told-to essay. He recommends treating chatbots like public postcards, using enterprise-grade models for work, regularly deleting chat history, and choosing well-known tools with clear privacy settings. These practices aim to reduce training leakage, long-term memory exposures, and data-breach risks.
Key Points
- 1Advocates treating AI chats like public postcards to avoid sharing personally identifiable information
- 2Warns that public models may train on conversations, causing training leakage and potential data exposure
- 3Recommends enterprise models, temporary chats, and regular history deletion to minimize leak and breach risk
Scoring Rationale
Practical, actionable privacy guidance from a Google engineer, but limited novelty and single-source reporting reduces breadth.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems