CANADALAND Could OpenAI Have Stopped Tumbler Ridge?
16 snips
Feb 25, 2026 Luke Savage, freelance journalist who writes for outlets like the Atlantic and Guardian, discusses OpenAI's handling of troubling ChatGPT interactions tied to the Tumbler Ridge shooting. He unpacks why US outlets break Canadian scoops. They examine what OpenAI reportedly saw and ignored, political reactions in Canada, and tensions between Silicon Valley's culture and safety obligations.
AI Snips
Chapters
Transcript
Episode notes
OpenAI Staff Debated Reporting Tumbler Ridge Warnings
- OpenAI flagged Jesse Van Rootselaar's ChatGPT interactions in June but decided they didn't meet the threshold to notify law enforcement.
- About a dozen OpenAI staff debated reporting after automated reviews flagged violent scenarios described over several days, per the Wall Street Journal.
Require Clear Reporting Thresholds From AI Firms
- Demand transparent, enforceable reporting thresholds for AI companies to balance safety and privacy.
- Luke Savage insists we need to know OpenAI's standards and what would have qualified as a reportable imminent risk.
Chatbots Can Actively Fuel Delusions
- Chatbots differ from passive media because they engage users and can simulate intimacy, increasing risk for vulnerable people.
- Luke Savage cites cases where prolonged chatbot exchanges produced delusions and spirals like Alan Brooks' 300-hour exchange that fueled conspiracy beliefs.