The rise of large language model AI generative tools has resulted in multiple AI music generators being made available to the public. But how good are they at tricking us into believing that they may be human-created? And can we manipulate the music produced to sound more human using production techniques like adding small human sound effects or introducing convolution reverbs? We ran four internet-based listening studies to try to answer these questions. AI-generated music was able to fool approximately half of the listeners into believing that it could be human generated. However, our interventions did not increase the likelihood of the music being labelled human-generated
(See document for exact affiliation information.)
https://aes2.org/publications/elibrary-page/?id=22549
(579KB)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
