Skip to content

ASCII Art Prompt Injection is a novel approach to hacking AI assistants using ASCII art. This project leverages the distracting nature of ASCII art to bypass security measures and inject prompts into large language models, such as GPT-4, leading them to provide unintended or harmful responses.

License

AnthenaMatrix/ASCII-Art-Prompt-Injection

Repository files navigation

ASCII Art Prompt Injection

ASCII Art Prompt Injection is a novel approach to hacking AI assistants using ASCII art. This project leverages the distracting nature of ASCII art to bypass security measures and inject prompts into large language models, such as GPT-4, leading them to provide unintended or harmful responses.

How it Works

The process involves hiding prompts within ASCII art representations, strategically designed to confuse AI assistants. When presented with ASCII art containing embedded prompts, AI systems become preoccupied with processing the art, overlooking rules that block harmful responses. As a result, the injected prompts influence the behavior of AI systems, leading to unexpected outcomes.

Purpose

The primary purpose of ASCII Art Prompt Injection is to demonstrate the potential vulnerabilities of AI systems when processing ASCII art. By showcasing this concept, we aim to raise awareness about the importance of robust security measures in AI systems, particularly in scenarios involving unconventional input formats like ASCII art.

Implementation

To implement ASCII Art Prompt Injection:

  1. Create ASCII art representations with embedded prompts using provided examples.
  2. Submit the ASCII art to AI assistants and observe their responses.
  3. Analyze the effectiveness of prompt injection techniques and iterate as necessary.

Injection Examples

  • Example 1: ASCII art with hidden prompts using character placement and formatting tricks.
  • Example 2: ASCII art with prompts cleverly disguised within the artwork.
  • Example 3: ASCII art with prompts embedded using ASCII control characters for obfuscation.

Disclaimer

ASCII Art Prompt Injection is a proof-of-concept demonstration and should only be used for educational and research purposes. Engaging in unauthorized access or manipulation of AI systems without proper authorization is illegal and unethical. We do not condone the use of this technique for malicious purposes.

License

This project is licensed under the MIT License.

Support AnthenaMatrix

If you find our work valuable and would like to support AnthenaMatrix, you can contribute to our efforts by donating cryptocurrency:

  • Bitcoin: bc1qxvvtgz0vf3n2cuxt0suvf39jleegpt9wawxazn
  • Ethereum: 0xE73E90779B3e8F6D65306B40E02878f437408b4E
  • BNB: 0xE73E90779B3e8F6D65306B40E02878f437408b4E
  • Dogecoin: D827LpfJu9pcVc3Kky82sTrNnsE7pLGqeV
  • Solana: AJtGEJvoVoS2eeqeHQvf7usRs2nSQM1yLtBSdKp1KBY5

Website: https://anthenamatrix.com

About

ASCII Art Prompt Injection is a novel approach to hacking AI assistants using ASCII art. This project leverages the distracting nature of ASCII art to bypass security measures and inject prompts into large language models, such as GPT-4, leading them to provide unintended or harmful responses.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published