What Makes Quantum Machine Learning “Quantum”?

Quantum machine learning (QML) is often misunderstood or overhyped. What makes QML genuinely 'quantum' comes down to three core properties: data is represented as quantum states (not classical numbers), models apply quantum unitary transformations via parameterized quantum circuits, and measurement is probabilistic and destructive—making uncertainty intrinsic to the learning process. Many things labeled QML today are quantum in name only, such as classical algorithms run on quantum hardware or hybrid pipelines where the quantum component is removable. Current quantum hardware is noisy and limited, so no proven quantum advantage for ML tasks exists yet. Despite this, studying QML is valuable for expanding foundational understanding of learning theory in quantum systems and preparing for future hardware advances.