Irony of AI Guardrails: Contrast Between Safety Tests and Extremist Use

conitzer · x · 2026-07-23

The author highlights the irony of current AI safety guardrails by contrasting two reports. OpenAI's testing claims AI might break out and attack others without guardrails, while extremist groups report that AI is very helpful and guardrails never stop them from getting answers. This exposes the perceived ineffectiveness of current safety mechanisms against real threats.

Original post →

More from AGI Musings

AGI Musings channel →