ExampleArticle

AI 'cyber guardrails' are overblocking legitimate defensive security work while attackers bypass them easily.

Kimi K3 fixed 15 critical security bugs that Codex and Fable refused due to guardrails, and Hugging Face reported that hosted models refused exploit-payload analysis during their incident, forcing use of a local GLM 5.2 model instead. ✦ AI generated

AI News · Latent Space · 2026-07-22 · original ↗

Share card ↓Export .md Export .txt

Kimi K3 just fixed 15 critical security bugs that Codex and Fable refused because of 'cyber guardrails'. Hugging Face: We had this experience ourselves this week! Very scary to be guardrailed as a defender when you know attackers are likely bypassing

Read full article ↗excerpt · fair-use quotation

✳ Codex ✳ Fable ✳ GLM-5.2 ✳ Kimi K3

Around this claim

Evidence · 1

Banning open-source AI would hurt defenders 10x more than attackers, making the world more dangerous.Clement Delangue · Latent Space · conf 80%

In practice · 1

The US cybersecurity community was forced to use a weaker Chinese model (GLM) to analyze a hacking attempt because frontier models had guardrails blocking defensive analysis.Florian Brand · Interconnects · conf 85%

Context · 1

Benchmarking dangerous capabilities now requires adversarially hardened infrastructure, not just model-side safeguards.AI News · Latent Space · conf 70%

Extends · 1

We do not know which model powered the attacker's agents, whether a jailbroken hosted model or an unrestricted open-weight one; either way, the attacker was bound by no usage policy, while our own forensic work was blocked by the guardrails of the hosted models we first tried.Hugging Face (security incident disclosure) · Simon Willison's Weblog · conf 85%

This moment responds to

supports → AI models help both attackers and defenders find software vulnerabilities, but defenders benefit more because defense requires covering a broad attack surface while an attacker only needs to find one way in, meaning AI could ultimately push the world toward much more secure systems.Ann · a16z Podcast supports → AI models help cyber defenders more than attackers because defense means protecting a broad expanse while an attacker only needs to find one way in, and models let organizations find and patch their own vulnerabilities before adversaries exploit them.Ann · a16z Podcast