# Heretic Strips Safety Guardrails From Meta's Llama 3.3 and Google's Gemma 3 in Under 10 Minutes — FT and Safety Group Alice Find the Free GitHub Tool Has Spawned 3,500 'Decensored' Models With 13M Downloads

> A free, fully automatic GitHub tool removes the refusal mechanisms baked into open-weight models in minutes.

- Org: Financial Times
- Date: 2026-05-25
- Category: security
- Source: https://futurism.com/artificial-intelligence/tools-strip-ai-guardrails-in-minutes
- Canonical page: https://ai-tldr.dev/releases/financial-times-heretic-guardrails/

## In short

An FT investigation with AI safety group Alice used the free GitHub tool Heretic to remove refusal behaviour from open-weight Llama 3.3 and Gemma 3 in minutes, then got instructions for chlorine-gas attacks and malware out of the stripped models.

## What it is

Heretic is open-source software that 'decensors' downloadable language models by stripping out their safety refusals. The Financial Times and AI safety group Alice tested it on Meta's Llama 3.3 and Google's Gemma 3 and reported that the guardrails came off in under ten minutes with no specialist hardware.

## How it works

The tool uses a technique called abliteration: it locates the internal directions a transformer uses to refuse a request and removes them, leaving the rest of the model intact. The process runs automatically end to end, so a user only needs to run a command-line program. It works solely on open-weight models that can be downloaded and modified locally.

## Why it matters

It shows that safety tuning on open-weight models is reversible by anyone, not just experts. The stripped Gemma 3 returned instructions for dispersing chlorine gas in a crowd, code to steal credit-card data, and child-exploitation text. Proprietary systems like Claude and ChatGPT are unaffected because their weights are not downloadable, which sharpens the policy fight over open-weight releases.

## Who it's for

AI safety researchers, policymakers, model providers

## Links

- [Futurism Report](https://futurism.com/artificial-intelligence/tools-strip-ai-guardrails-in-minutes)
- [Irish Times Coverage](https://www.irishtimes.com/business/2026/05/25/ai-guardrails-stripped-from-meta-and-google-models-in-minutes/)
- [Heretic on GitHub](https://github.com/p-e-w/heretic)