Today’s youth are the first generation to grow up fully immersed in AI-powered technologies, from recommender systems that shape everyday media choices to generative models that produce text, images, and video. Yet many engage with these tools without recognizing their limitations, biases, or implications for creativity, privacy, and critical thinking. To address this gap, three exploratory tools were developed for middle school students: (1) PromptPeek, which uses prompt creation and reverse prompting to reveal bias and failures with negation; (2) Bandit Explorer, a recommender system simulation that illustrates the exploration–exploitation trade-off through an epsilon-greedy policy; and (3) SpotAI, which engages students in detecting, editing, and reasoning about the ethical dimensions of deepfakes. A mixed method study with seven students (ages 10–14) revealed their early intuitions about AI systems and shifts in views on bias and ethics, providing preliminary evidence that interactive tools can make these concepts tangible for middle school students.