Grade 8: AI and Human Oversight
Students deepen their understanding of AI by looking beyond what AI can do to investigate what resources it depends on, and how it affects people and society
- Teachers
- Curriculum & Resources
- Grade 8: AI and Human Oversight
Age:
AI and Human Oversight
Approximate Total Time: 5 class periods (approximately 50 minutes each)
Recommended for Ages 13–14 (typically U.S. Grade 8)
About This Unit
Students deepen their understanding of artificial intelligence by looking beyond what AI can do to investigate how it works, what resources it depends on, and how it affects people and society. Students compare generative and discriminative AI systems, explore the environmental costs associated with AI, and investigate how recommendation systems use behavioral data to personalize what people see and potentially shape attention, choices, and perspectives.
Students then examine how bias can enter AI systems through training data, labels, shortcuts, testing, and past patterns of advantage or disadvantage. In the final lesson, students act as human reviewers who must decide whether to approve, verify, or override an AI prediction or recommendation. Throughout the unit, students weigh AI's benefits, limitations, tradeoffs, and consequences while recognizing the importance of human judgment.
Lessons
1. Machine Learning and Its Many Forms
Students compare generative and discriminative AI systems and investigate how different machine-learning systems learn patterns from data.
2. The Real Cost of AI
Students investigate the electricity, water, hardware, infrastructure, and other resources required to support AI and evaluate environmental tradeoffs associated with different uses.
3. AI, Power, and Persuasion
Students explore how recommendation systems use behavioral data to decide what people see and analyze how human choices about optimization can shape attention, choices, and perspectives.
4. Who Gets Left Out By AI?
Students investigate how bias can enter AI systems through data, labels, shortcuts, and testing and analyze real-world examples to identify what happened, who was affected, and possible causes.
5. When Should Humans Override AI?
Students act as human reviewers and decide whether to approve, verify, or override AI decisions in different scenarios while developing criteria for when human judgment should take priority.