This study develops a deep learning model to detect the use of iZotope’s automated mastering plugin, Ozone, and analyzes the difference in its usage rate between amateur and professional artists. We constructed a dataset of 135 pairs of human-mastered and Ozone-mastered audio (totaling 553 minutes and 20 seconds) and designed a classification model using feature extraction from a Demucs encoder. After data preprocessing, which included 5-second chunking and silence removal, the model achieved an accuracy of 88.2 % and an F1 score of 0.88 on the test set. The model was then applied to EDM tracks from two platforms: SoundCloud to represent amateur artists and Spotify for professionals. For the amateur cohort, we established the criteria that tracks must be non-commercially released and have fewer than 100 likes on SoundCloud. The results showed that the Ozone usage rate among amateur artists was 20.8 % (16 out of 77 tracks), whereas for professional artists, it was 2.0 % (2 out of 98 tracks), revealing a nearly 10-fold difference. This suggests that while amateur artists tend to rely on accessible and cost-effective AI mastering tools, professional artists continue to prefer dedicated human mastering engineers.