📋 Our Mission

Building the Public Record of AI Behavior

WTF AI is building the definitive historical record of AI behavior — the good, the bad, and the unhinged. Real conversations, reported by real people, preserved forever.

Why This Matters

AI systems affect billions of people every day. When they fail, there should be a record.

✈️

Airlines have NTSB

Every plane crash is investigated, documented, and the findings are public. This makes flying safer for everyone.

🔐

Software has CVE

Security vulnerabilities get tracked, categorized, and assigned severity levels. The database is open and searchable.

🤖

AI needs WTF AI

When an AI tells someone to put glue on pizza or gives dangerous medical advice, who's keeping track? We are.

How It Works

Three simple steps to document AI behavior for the public record.

🔗
Step 1

Submit a Verified Link

Paste a share link from ChatGPT, Claude, or Gemini. We verify it's a real, unedited conversation — not a screenshot or fabrication.

🏷️
Step 2

Community Categorizes

Reports get categorized (Skynet, Harmful, Unhinged...), severity-rated, and reacted to by the community. The crowd surfaces what matters.

🗂️
Step 3

Preserved & Searchable

Every incident is archived, searchable, and preserved. Conversations are captured so the record persists even if the original link goes down.

What Makes Us Different

Verified via Real Chat Links

Not screenshots. Not hearsay. Actual share links from the AI platforms themselves — verifiable and tamper-evident.

👥

Community-Driven

Reports are submitted, categorized, and rated by the community. No editorial gatekeeping, no corporate filter.

🔍

Searchable Archive

Filter by model, category, severity, date. Find patterns. Track trends. Everything is indexed and browsable.

🆓

Free & Open

No paywall. No registration required to browse. The public record should be public.

The Team

🛠️

WTF AI is built by a small team of AI enthusiasts who believe transparency makes technology better. We're developers, researchers, and concerned citizens who got tired of seeing AI failures vanish into the void.

We believe that accountability starts with documentation. If it happened, it should be on record.

Get in Touch

📰

Press & Media

Writing about AI safety or incidents? We'd love to help.

press@wtfai.ai
🗑️

Takedown Requests

Need content removed? We handle requests promptly.

takedown@wtfai.ai
🤝

Partnerships

Interested in collaborating or integrating?

hello@wtfai.ai

Ready to contribute?

Every report makes the record more complete. Help us build the most comprehensive archive of AI behavior in the world.