This repository curates examples of "awful AI" – AI applications exhibiting bias, misuse, or harmful societal impacts. It serves researchers, engineers, and the public by documenting concerning AI trends and fostering discussion on ethical development and potential countermeasures.

How It Works

The project categorizes documented AI misuses into distinct areas like discrimination, disinformation, surveillance, and autonomous weapons. Each entry provides a summary, detailed explanation, and references to original sources, allowing for a comprehensive understanding of specific AI failures and their real-world consequences.

Quick Start & Requirements

This is a curated list, not a software project. No installation or execution is required. Access is via web browser.

Highlighted Details

• Comprehensive categorization of AI harms, including discrimination, disinformation, surveillance, data crimes, social credit, scams, climate impact, and autonomous weapons.
• Detailed case studies with references, such as the COMPAS algorithm's racial bias, Tay chatbot's racism, and the use of AI in warfare.
• Includes a "Contestational AI Efforts" section highlighting projects and research aimed at mitigating AI harms.
• Features an annual "Awful AI Award" to recognize particularly egregious examples of unethical AI development or deployment.

Maintenance & Community

The project is maintained by David Dao. Community engagement and discussion are encouraged through the repository's issue tracker.

Licensing & Compatibility

David Dao has waived all copyright and related or neighboring rights to this work, effectively placing it in the public domain.

Limitations & Caveats

The list is a curated collection and does not represent an exhaustive catalog of all AI misuses. The information is presented as documented in the cited sources.