Choose an appropriate AI approach for a task by comparing how different AI systems operate and what they are best suited to do.

What you learn at this level

• Learners recognise that some types of AI systems can be programmed with specific rules to accomplish tasks, while others can learn patterns from data.

• Learners compare different approaches to building AI systems by completing a trash-sorting task. First, they consider why random outputs would be unlikely to reliably sort items. Next, they direct a classmate role-playing as a robot to sort images of trash based on specific criteria. They observe how the “robot” responds to new images that do not match the criteria. Finally, they explore how a system could use patterns from collected data, such as a class survey, to improve sorting decisions and determine which approach is most appropriate for different tasks.

What you learn at this level

• Learners identify the benefits and drawbacks of using a rules-based approach and a machine learning approach to solve a problem.

• Learners compare technology that follows set rules to execute a task (e.g. calculators, traffic lights, thermostats) with technologies that have been trained on examples from data (e.g., image recognition, translation). With a teacher, they consider why each approach suits the specific task.

What you learn at this level

• Learners evaluate when AI is an effective approach for a task by considering factors such as context, data availability and quality, efficiency, transparency, desired outcomes and potential impacts.

• Learners analyse real-world examples of AI use (e.g. route planning, content moderation or resource allocation). They justify whether a rule-based or machine learning approach is more appropriate – or whether AI should be used at all – based on the requirements, constraints and trade-offs involved. They build a case to make a recommendation, then present their chosen approaches to their peers in a class debate.