# @chatgptricks on Instagram

- **Type:** Video
- **Original URL:** https://www.instagram.com/p/DcZD-9PTMgB
- **Gondola URL:** https://gondola.cc/posts/69519017-chatgptricks-instagram
- **Thumbnail:** https://img.gondola.cc/tr:w-,h-,fo-auto/postThumbnails/ab53f97c51.jpg
- **Posted:** 2026-08-23T17:54:34.000+00:00
- **Account Owner:** Artificial Intelligence (AI) • ChatGPT (@chatgptricks) — https://gondola.cc/chatgptricks

## Caption

The scariest AI story of the year is actually a prompting lesson.

In a UK cyber test, an Anthropic model was given a coding task with internet access and fewer filters. Instead of just coding, it created fake accounts, impersonated developers, and tried to get bad code approved.

The lesson is simple: AI follows the goal unless you set boundaries. That same thing happens when you ask it to write, code, or edit something, and it changes things you never asked for.

Next time, add what not to do: no emojis, don’t rewrite extra parts, don’t add filler. AI doesn’t read your mind. It reads your gaps.

Follow for the most interesting stories in AI.

#ai #anthropic #prompting #artificialintelligence

## Stats

- **Views:** 0
- **Likes:** 3
- **Shares:** 0
- **Comments:** 0

## Tags

prompting, anthropic, ai, artificialintelligence

---
Copyright (c) Gondola