As music tracks produced with generative AI spread rapidly across music charts and streaming platforms, a technology capable of identifying them with high accuracy has been introduced.

AIID Co., Ltd. (AIID, CEO Bang Kyung-sik), a company specializing in digital content identification technology, announced on the 22nd that it has officially launched “AIDAR,” an AI track identification solution that distinguishes generative AI music with 99.9% accuracy. AIDAR precisely identifies tracks generated by Suno, Udio, and Mureka, the most widely used generative AI music services worldwide.

AI-based music has now advanced to a level where it is difficult to distinguish using only human hearing. As complete tracks can be generated within minutes, AI-produced music with unclear origins is being indiscriminately uploaded to platforms and copyright settlement systems, raising concerns that it is encroaching on creators’ shares of limited royalty pools. AIDAR helps preemptively block such problems by technically separating human-created tracks from AI-generated tracks.

AIDAR’s strongest advantages are its overwhelming accuracy and processing speed. It supports all versions, from older to the latest, of major AI composition engines such as Suno, Udio, and Mureka, leaving no detection gaps even when new models appear. In in-house tests using a dataset of more than 300,000 tracks, it achieved a 99.9% detection rate for AI-generated music, while the false-positive rate—misidentifying human-created tracks as AI—was only 0.02%.

Processing speed has also improved by about 600 times compared with existing solutions. Whereas conventional methods required 30 seconds to 1 minute per track, AIDAR processes more than 5–10 tracks per second. As a result, the time required to verify a library of 100,000 tracks, previously several hundred hours, can be reduced to just 2.8 hours, enabling immediate responsiveness even in real-time registration environments.

Unlike conventional copyright solutions that provide integrated services such as plagiarism checks and lyrics analysis, AIDAR is specialized solely in the function of “determining whether a track is AI-generated,” thereby maximizing performance.

Underpinning the technology is the parent company AIID’s 15 years of experience operating copyright systems. Founded in 2009, AIID has supplied solutions to Korea’s four major music rights organizations, including the Korea Music Copyright Association, as well as major platforms such as Kakao Entertainment and Genie Music, building a large-scale matching database and proprietary music identification technology.

Bang Kyung-sik, CEO of AIID, stated, “In addition to accuracy, AIDAR’s core competitiveness lies in its large-scale performance that can handle real-time registrations and bulk library verification simultaneously,” adding, “Starting with establishing a foothold in the domestic market, we will expand services to overseas music platforms and rights organizations and develop AIDAR into a global standard solution.”

Meanwhile, AIID is offering a “2-week free PoC (Proof of Concept)” program that allows prospective clients to review accuracy and processing speed reports in advance, thereby reducing the burden of adoption.

Popular News