Shipped: AI Security Literacy (AISL) — Research & Benchmarking Framework

AI Security Literacy (AISL) is an open-source research and benchmarking framework designed to make AI security education measurable, reproducible, and accessible. It defines six competency dimensions spanning privacy, threat recognition, trust calibration, adversarial AI, responsible use, and learner agency.

Built as a reproducible Google Colab pipeline, AISL combines public AI benchmarks with national survey data to investigate real-world AI risks. Experiments include hallucination and misinformation, toxicity detection, scam classification, social bias, and adversarial attacks using FGSM and PGD.

The project provides reusable notebooks, model evaluations, visualizations, and research outputs to help educators, researchers, and developers understand how AI systems fail and how AI security literacy can be taught and assessed.