This repository is a curated compilation of resources on Offensive AI, targeting security researchers, AI developers, and ethical hackers. It aims to provide a comprehensive overview of AI vulnerabilities, attack vectors, and defensive strategies, enabling users to understand and mitigate risks associated with AI systems.

How It Works

The compilation is structured into categories covering various aspects of Offensive AI, including Adversarial Machine Learning (attacks like extraction, inversion, poisoning, evasion), AI-powered malware, OSINT, phishing, and the misuse of generative AI across audio, image, video, and text modalities. It links to research papers, tools, and practical applications for each category.

Quick Start & Requirements

• No installation or execution required; this is a curated list of links and resources.
• Access to the internet is the primary requirement.
• Links to official documentation, demos, and related projects are provided within the repository.

Highlighted Details

• Extensive coverage of adversarial attacks: extraction, inversion, poisoning, and evasion, with detailed explanations and numerous linked resources.
• Broad scope of generative AI misuse: includes tools and applications for audio, image, video, and text manipulation and detection.
• Practical application examples: showcases AI in pentesting, malware development, OSINT, and phishing.
• Tooling and framework support: lists popular libraries like Cleverhans, ART, Foolbox, and TextAttack for adversarial machine learning.

Maintenance & Community

• Maintained by Miguel Hernández and José Ignacio Escribano.
• Community engagement details (e.g., Discord/Slack links) are not specified in the README.

Licensing & Compatibility

• License: Creative Commons Attribution-Share Alike 4.0 International.
• This license allows for sharing and adaptation, provided attribution is given and any derivative works are shared under the same license. It is generally compatible with commercial use and closed-source linking.

Limitations & Caveats

The repository is a static compilation of links and does not offer active tools or services. The rapidly evolving nature of AI security means some linked resources may become outdated.