Malaysian music publishers are monitoring whether locally produced songs were used without permission to train generative AI models, following an investigation by _The Atlantic_ journalist Alex Reisner that uncovered four massive datasets containing more than 21 million songs. _The Atlantic_ launched The AI Watchdog, a search engine allowing users to check if works appear in the datasets. A check by Malay Mail found at least 1,000 songs by Malaysian artistes in the datasets, including 106 songs by Datuk Seri Siti Nurhaliza, 134 songs by the late Tan Sri P. Ramlee, 151 songs by Yuna, and 73 songs by Butterfingers. At least one dataset contains links to music on YouTube and Spotify. _The Atlantic_ stressed that a song’s presence in a dataset is not definitive proof it was used to train an AI model. Taja Archive Sdn Bhd Malaysia director Kimmy Khir said the publisher is monitoring developments and will assess action if there is credible evidence of unauthorised use, adding that the industry needs greater transparency, clearer licensing practices, and meaningful engagement between AI developers and rights holders.