AI music solutions company Muzig AI said on Feb. 26 it has developed 'Sound AI Detecting' technology that can precisely distinguish between music it generates and music created by humans.

The company said the 'Sound AI Detecting' technology it developed conducts combined analysis of a wide range of elements. It is based on deep learning and machine learning algorithms and examines not only the acoustic characteristics of music but also melody development patterns and structural regularities in rhythm.

While existing analysis techniques simply compared sound waveforms, the company said the new technology traces back the source of music through randomness that only AI-generated output has, a characteristic of AI technology. It said this allows it to accurately distinguish digital patterns unique to AI models from the irregular characteristics unique to human-created works.

Muzig AI Chief Technology Officer Moon-bong Choi (최문봉) said, "As AI music and other generation technologies become more advanced, the importance of technology that clearly identifies their origin and rights relationships will grow further." He added, "I hope this technology raises transparency for content generated by AI and becomes a standard technology that protects creators' rights and interests."

Muzig AI said it is speeding up efforts to build a deepfake voice tracking system based on the sound tracking programme it has already completed development of. It said it plans to continue rolling out solutions to secure technological competitiveness in the global music market and to protect AI content ethics and copyright.