Certified Tester Performance Testing (CT-PT)

A test manager is evaluating the CT-AI Version 2.0 syllabus and notices that a colleague expects it to still cover using AI tools to assist their own test execution work. Based on the syllabus scope, how should the manager correct this expectation?

A tester is preparing to test a machine-learning-based system and struggles because expected outputs cannot be defined precisely and behavior varies across runs. Which specific testing challenges, as described for AI-based systems, best explain the tester's difficulty?