AI HORRORS
Community Driven

Submit Your
AI Horror Story

Help the community learn from real AI disasters. Share your production failures, security breaches, and cautionary tales.

⚠

Submission Guidelines

Before submitting, ensure your story meets these requirements:

  • Real incidents only – No fiction or hypothetical scenarios
  • Remove sensitive data – Credentials, PII, internal URLs must be redacted
  • Anonymize if needed – Company names optional (or get permission first)
  • Include lessons learned – Every story should have actionable takeaways

How to Submit

1

Fork the Repository

Start by forking the AI HORRORS repository on GitHub

Fork on GitHub
2

Create Your Story File

Add a new markdown file to the content/blogs/ directory using our template

---
title: "Your Story Title"
date: "2026-05-01"
severity: "critical"  # critical | high | medium
tags: ["database", "production", "security"]
excerpt: "Brief 1-2 sentence summary (max 150 chars)"
---

## What Happened

Describe the incident timeline and events...

## The Impact

- Downtime duration
- Data loss details
- Financial cost
- Users affected

## How It Was Fixed

Step-by-step resolution process...

## Lessons Learned

Key takeaways and prevention strategies...
3

Submit a Pull Request

Create a PR with your story and we'll review it promptly

Create Pull RequestView Full Guide

Template Fields

Frontmatter (YAML)

Metadata at the top of your markdown file

  • titleCatchy, descriptive title that hints at the disaster
  • dateWhen the incident occurred (YYYY-MM-DD format)
  • severitycritical | high | medium
  • tagsArray of relevant tags (e.g., ["database", "llm", "production"])
  • excerptBrief summary for card preview (max 150 characters)

Severity Levels

Choose the appropriate severity for your incident

🔴Critical

Production outage, data loss, security breach, major financial impact

🟠High

Significant disruption, contained damage, notable incident response

🟡Medium

Recoverable failure, limited impact, valuable learning opportunity

💡

Pro Tips

  • ▸Use descriptive titles that capture the "oh no" moment
  • ▸Include code snippets where relevant (properly escaped)
  • ▸Link to official incident reports, tweets, or blog posts if available
  • ▸Screenshots are great – add them to public/images/
  • ▸Focus on lessons learned – that's what makes stories valuable

Need Help?

For detailed guidelines, check out our contribution guide or browse existing stories for reference.

Contributing GuideBrowse Examples