AI-enabled fraud at scale exploiting royalty pool economics: synthetic content dilution reduces legitimate artist compensation

str 8 3/18/2026 · 1 article

economic · technological · AI, Media · FR, Global

Analysis

Fraudsters deploy AI generation tools to rapidly create thousands of tracks and bot networks to game streaming algorithms and extract royalties. When fraudulent tracks are removed from shared royalty pools, the per-stream payout for legitimate artists declines, creating a structural vulnerability where platform growth in content volume directly harms creator economics and enables profitable fraud at scale.

Key actors

DeezerIFPI

Source article

French music streamer Deezer battles deluge of AI fraud

Financial Times — AI, Data, Robotics and Digital Power · 3/18/2026 · extracted in run pdf-import-2026-03-18-1779256502138-57 · 5/20/2026, 6:45:40 AM

"fraudsters were responsible for more than 80 per cent of all streams of AI-generated music" [80 per cent]

The 80% figure directly quantifies the scale of fraud within AI-generated content, establishing that the majority of AI streams on the platform are fraudulent rather than legitimate creative use.

"When tracks are detected as fraudulent they are removed from the royalty pool, which is shared among all artists and songwriters on Deezer's platform" [royalty pool]

This mechanism shows that fraudulent tracks, even when removed, have already diluted the pool during detection lag, and removal itself doesn't restore lost payouts to legitimate artists.

Reasoning from this article

The article reveals that AI generation tools (Suno, Udio) enable rapid, low-cost track creation, while streaming economics allow fraudsters to accumulate revenue through volume. The 60,000+ AI tracks added daily (39% of daily intake) and 85% fraud rate on AI content shows this is not marginal—it's becoming the dominant use case for AI in music streaming. This pattern will likely spread across all platforms as tools democratize and detection lags.

The article indicates fraudulent streams comprise ~8% of all Deezer streams in 2025, meaning legitimate artists' per-stream payouts are suppressed by this volume. As AI track uploads accelerate (60,000+ daily), the detection-to-removal lag creates a window where fraud actively extracts value from the shared pool. This creates pressure on platforms to either tighten payout economics or implement stricter content gatekeeping.