The Right to Red-Team: Adversarial AI Literacy as a Civic Imperative in K-12 Education

Devan Walton, Haesol Bae

Keywords: AI literacy, K-12 education, Red-teaming, AI safety

TL;DR: Schools should teach kids to hack AI systems (safely) so they can hold them accountable as citizens.

Abstract: The increasing societal integration of Large Language Models (LLMs) and agent-based AI demands a new civic competency: adversarial reasoning. This position paper argues that K-12 AI education must move beyond passive literacy to actively equip students with skills in responsible adversarial prompting and ethical system "hacking." Such capabilities are essential for citizens to critically probe AI systems, understand their inherent limitations, identify manipulative patterns, and hold them accountable. We posit that cultivating a generation skilled in "red-teaming" AI is vital for maintaining transparency, preventing undue influence, and fostering a democratic engagement with these transformative technologies.

Lay Summary: AI systems like ChatGPT now help millions of people with homework, work tasks, and daily questions. But these systems regularly fail in dangerous ways. Simple word tricks can make them give harmful advice, spread false information, or reveal private data. Companies try to fix these problems internally, but they cannot catch every failure before it reaches users. We argue that students should learn how to test and find weaknesses in AI systems as part of their regular K-12 education. Just as students learn to fact-check sources and understand media bias, they need skills to probe AI systems and expose their failures. We call this adversarial reasoning. Our paper shows that even elementary students can learn these skills in just a few hours using simple activities. Teaching these techniques makes students more responsible rather than encouraging misuse. We propose concrete policy changes. First, we suggest adding AI testing to existing computer science standards. Second, we recommend legal protections for students who responsibly report AI flaws. Third, we propose twenty-five million dollars in federal funding so every U.S. high school can teach these skills. When AI systems shape our information and decisions, citizens need the power to question them.

Submission Number: 428