Fraudster Used AI Songs and 10K Bots to Steal Royalties

In what prosecutors describe as one of the first criminal cases of its kind, a North Carolina musician has admitted to orchestrating a massive streaming fraud scheme powered by AI-generated music and an army of bots. The case offers a striking window into how generative audio tools, when weaponized at scale, can quietly siphon millions from royalty pools meant for human artists.

According to the Department of Justice, Michael Smith pleaded guilty to charges stemming from a scheme that used hundreds of thousands of AI-generated songs and roughly 10,000 automated bot accounts to game streaming platforms including Spotify, Apple Music, and Amazon Music. Over the course of the operation, Smith allegedly collected more than $10 million in royalty payments — fraudulent streams that, as the headline notes, could have 'outstreamed Taylor Swift' in sheer volume.

How the Scheme Worked

The mechanics of the fraud reveal how cheap and scalable AI music generation has become. Rather than relying on a small catalog of real tracks, Smith reportedly partnered with the CEO of an AI music company and a music promoter to generate an enormous library of synthetic songs. By spreading fake streams across hundreds of thousands of tracks rather than concentrating them on a handful, the scheme aimed to stay under the radar of anti-fraud detection systems.

Streaming platforms typically flag suspicious activity when a single track accumulates implausible play counts. But distributing roughly 660,000 daily streams across a vast catalog of AI-generated filler made each individual song's numbers look modest and organic. The bots cycled through these tracks continuously, each stream triggering a micro-payment that, aggregated across years and accounts, added up to a seven-figure payday.

Why AI Music Made This Possible

The scheme highlights a structural vulnerability in the economics of streaming. Royalty pools distribute payouts based on a platform's total stream share, meaning fraudulent plays effectively steal from legitimate artists by diluting the pool. What makes this case distinct from earlier streaming fraud is the role of generative AI: producing hundreds of thousands of unique-sounding tracks once required studios, musicians, and time. Now, text-to-music models can spin up endless catalogs on demand, at near-zero marginal cost.

This collapses the barrier to entry for industrial-scale fraud. A bad actor no longer needs to license or produce real music — they can synthesize a bottomless inventory of plausible songs, each serving as a container to absorb bot-driven streams. The AI-generated nature of the catalog also makes takedowns and attribution harder, since there's no human creator or recognizable source to trace.

A Test Case for Synthetic Media Enforcement

The guilty plea marks an important precedent for how the legal system handles fraud enabled by synthetic media. Smith faces charges including wire fraud and money laundering conspiracy, each carrying substantial prison time. Prosecutors framed the case as a theft from real musicians and songwriters whose share of royalty pools was eroded by the fake streams.

For the broader synthetic media landscape, the case underscores a recurring theme: detection lags generation. Streaming platforms invest heavily in bot-detection and anomaly analysis, but fraudsters adapt by spreading activity thin and exploiting the sheer volume that AI enables. The same dynamic plays out across deepfake video, voice cloning, and AI-generated text — the tools that create synthetic content are advancing faster than the systems designed to catch misuse.

Implications for Platforms and Artists

Expect streaming services to tighten scrutiny on AI-generated uploads and high-volume catalog distributors. Some platforms have already begun experimenting with AI-content labeling and tighter payout thresholds for new or low-engagement tracks. The challenge is distinguishing legitimate AI-assisted music — an increasingly mainstream creative practice — from purpose-built fraud infrastructure.

Content authenticity and provenance tooling may become part of the answer. If platforms can reliably identify synthetic audio and correlate it with suspicious streaming patterns, they can throttle payouts to fraudulent catalogs before royalties are disbursed. But as this case shows, the economics strongly favor attackers until detection catches up.

The Smith case is likely the first of many. As AI music generation grows cheaper and more realistic, the incentive to exploit royalty systems will only intensify — making this guilty plea an early marker of a fraud category that regulators, platforms, and artists will be grappling with for years.

Stay informed on AI video and digital authenticity. Follow Skrew AI News.