UChicago launches Quicksilver tool to spot AI‑generated music

A new browser extension called Quicksilver, developed by University of Chicago researchers, can flag AI‑generated tracks while users stream music online [1]. The tool arrives as AI‑produced songs now account for a large share of new uploads, prompting concerns over transparency and the impact on human artists.

Key takeaways

• Quicksilver identifies AI‑made music in real time on a listener’s computer [1].
• In 2024 AI‑generated songs made up nearly half of all newly uploaded tracks on streaming services [1].
• Professional musicians perform only slightly better than random guessing at distinguishing AI from human music [1].
• The extension was created by the UChicago SAND Lab with support from the nonprofit Ethical Technology and Computing for Humanity [1].

How Quicksilver works and its development background

Quicksilver operates as a browser extension that analyzes audio streams and alerts users when a track is likely produced by artificial intelligence [1]. The project emerged from the University of Chicago’s SAND Lab—focused on security, algorithms, networking, and data—and involved Ethical Technology and Computing for Humanity (ETCH), a nonprofit launched by professors Ben Zhao and Heather Zheng [1]. Lead developer Stanley Wu, a graduate student, described the tool as a natural response to the “spammy” wave of AI music that threatens human creativity [1].

The extension follows earlier initiatives such as Glaze and Nightshade, which aim to disrupt non‑consensual AI training on artistic works [1]. By giving everyday listeners a way to detect AI‑generated songs, Quicksilver seeks to promote transparency in a market where AI music has surged. Researchers noted that even trained musicians struggle to tell AI tracks apart from human‑made ones, performing only marginally better than chance [1].

Context: AI music’s rapid rise and industry reaction

The explosion of AI‑generated music is part of a broader trend highlighted by recent industry reports. Suno, a leading AI music company, announced $300 million in annual recurring revenue and two million paying subscribers in early 2024, while claiming millions of users have tried its free tools [2]. Suno’s technology lets users generate songs from text prompts, upload audio, and even create AI versions of their own voices [2]. This growth has sparked legal disputes, with major labels accusing AI firms of training on copyrighted recordings without permission [2].

The popularity of AI‑created tracks is evident in chart performance as well. A Suno‑generated song titled “Celebrate Me” topped iTunes charts in several countries, including the United States, under the AI persona IngaRose [3]. While the identity behind IngaRose remains unclear, the song’s viral spread on —appearing in nearly 300 000 videos—illustrates how AI music can quickly capture public attention [3].