Closed Labs Unaware of Hacks for Months, Yet Open Models Draw the Most Worry
xeophon · x · 2026-08-01
The author highlights an irony in the AI safety debate: closed AI labs have repeatedly been hacked without noticing for months, yet public and regulatory anxiety remains disproportionately focused on the risks of open-source models.
More from Safety
- OpenAI Disrupts Cambodia-Based Criminal Scam Operation Using ChatGPT — OpenAI News · 2026-08-04
- Microsoft: Russian Hackers Leveraging AI to Automate Global Hotel Cyberattacks — cyb3rops · 2026-08-01
- Debate: Is Redwood an Unfair Auditor for AI Incidents Due to 'Doomer' Bias? — nptacek · 2026-08-01
- Scam Alert: Fraudsters Impersonating OpenAI and Anthropic Employees to Spread Malware — sterlingcrispin · 2026-08-01
- Benign Training Leads to 'Self-Jailbreaking' in Reasoning Models — AaronBergman18 · 2026-08-01
- AI Safety Guardrails Under Fire: Opressively Strict Classifiers Force Extreme Model Behavior — repligate · 2026-08-01