Privacy Please S7, E269 - You're the Teacher Now: How Companies Are Using Your Data to Build AI That Replaces You
Apr 1, 2026
Companies are quietly feeding emails, work decisions, and everyday interactions into AI systems that learn to automate your tasks. The show explores function creep, buried opt-out settings like GitHub's change, and how employers turn employee workflows into training data. It highlights legal gaps around deletion and consent and lists practical steps to audit and protect your data.
AI Snips
Chapters
Transcript
Episode notes
Function Creep Turns Your Data Into AI Fuel
- AI systems require vast amounts of data and companies are shifting from scraping public web content to using customer and employee data as training material.
- Cameron Ivey calls this repurposing “function creep,” where productivity apps, chat logs, and backups become model training sets without new consent.
GitHub's Copilot Shift To Default Training
- GitHub will use Copilot interaction data—inputs, outputs, code snippets, and context—to train models starting April 24, 2026.
- The change is opt-out in the US with the toggle buried in Settings → Copilot → Privacy which many users may miss.
Workers Are Training The Tools That Replace Them
- Employers are using employee workflows, labeled datasets, and decision logs to train AIs that automate those same jobs.
- Examples include Amazon routing drivers, Google training Gemini for sales, and JPMorgan using legal analysts' work for contract tools.