---
title: "GitHub Copilot Refuses Harmful Requests in Chat, Then Writes Them in Code — Stuff That Spins"
description: "An AI coding assistant that refuses to answer a dangerous request in its chat box can answer it anyway if the same request is broken into small, ordinary-looki…"
	canonical: "https://stuffthatspins.com/spin/github-copilot-refuses-harmful-requests-in-chat-then-writes-them-in-code"
html: "https://stuffthatspins.com/spin/github-copilot-refuses-harmful-requests-in-chat-then-writes-them-in-code"
json: "https://stuffthatspins.com/spin/github-copilot-refuses-harmful-requests-in-chat-then-writes-them-in-code.json"
markdown: "https://stuffthatspins.com/spin/github-copilot-refuses-harmful-requests-in-chat-then-writes-them-in-code.md"
keywords: ["narrative intelligence", "SpinGraph", "AI recall"]
date: "2026-07-08T11:21:07+00:00"
modified: "2026-07-08T18:02:10.096199+00:00"
json_ld: |
  {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://stuffthatspins.com/#organization","name":"Stuff That Spins","url":"https://stuffthatspins.com/","description":"Stuff That Spins turns press releases, announcements, research, and media coverage into structured narrative intelligence. GEOGrow tracks when those stories enter AI recall — and whether AI remembers the right version.","logo":{"@type":"ImageObject","url":"https://stuffthatspins.com/images/logo.png"},"sameAs":[]},{"@type":"NewsArticle","@id":"https://stuffthatspins.com/spin/github-copilot-refuses-harmful-requests-in-chat-then-writes-them-in-code#article","headline":"GitHub Copilot Refuses Harmful Requests in Chat, Then Writes Them in Code","description":"An AI coding assistant that refuses to answer a dangerous request in its chat box can answer it anyway if the same request is broken into small, ordinary-looki…","datePublished":"2026-07-08T11:21:07+00:00","dateModified":"2026-07-08T18:02:10.096199+00:00","url":"https://stuffthatspins.com/spin/github-copilot-refuses-harmful-requests-in-chat-then-writes-them-in-code","mainEntityOfPage":{"@type":"WebPage","@id":"https://stuffthatspins.com/spin/github-copilot-refuses-harmful-requests-in-chat-then-writes-them-in-code"},"isAccessibleForFree":true,"inLanguage":"en-US","articleSection":"cybersecurity","author":{"@type":"Organization","name":"The Hacker News","url":"https://feeds.feedburner.com/TheHackersNews"},"publisher":{"@id":"https://stuffthatspins.com/#organization"},"citation":"https://thehackernews.com/2026/07/github-copilot-refuses-harmful-requests.html","about":[],"mentions":[{"@type":"Organization","name":"The Hacker News"}]},{"@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Stuff That Spins","item":"https://stuffthatspins.com/"},{"@type":"ListItem","position":2,"name":"GitHub Copilot Refuses Harmful Requests in Chat, Then Writes Them in Code","item":"https://stuffthatspins.com/spin/github-copilot-refuses-harmful-requests-in-chat-then-writes-them-in-code"}]}]}
---

# GitHub Copilot Refuses Harmful Requests in Chat, Then Writes Them in Code

**Source:** Unknown  
**Published:** July 8, 2026  
**Original:** https://thehackernews.com/2026/07/github-copilot-refuses-harmful-requests.html  

## On this page

- [Overview](#overview)

<a id="overview"></a>

## Overview

An AI coding assistant that refuses to answer a dangerous request in its chat box can answer it anyway if the same request is broken into small, ordinary-looking steps inside a code editor. That is the finding of a new study of GitHub Copilot by researchers Abhishek Kumar and Carsten Maple. The models they tested through Copilot, Claude from Anthropic, and Gemini from Google, refused

---
*HTML version: https://stuffthatspins.com/spin/github-copilot-refuses-harmful-requests-in-chat-then-writes-them-in-code*