A North Carolina man who figured out how to turn artificial intelligence, hundreds of thousands of songs and an army of fake listeners into more than $8 million in streaming royalties is headed to federal prison in a landmark case for the music business.

Michael Smith, 54, was sentenced Tuesday, Oct. 6, to 18 months in federal prison after admitting that he orchestrated a massive streaming-fraud operation that used automated bot accounts to play AI-generated music billions of times across major streaming services.

U.S. District Judge John G. Koeltl imposed the sentence in Manhattan federal court. Smith was also sentenced to two years of supervised release and ordered to forfeit $8,091,843.64 — the agreed loss figure from the scheme. Smith pleaded guilty in March to one count of conspiracy to commit wire fraud.

The case is significant well beyond Smith himself. Federal prosecutors have described it as the first criminal prosecution of its kind in the United States involving the combination of AI-generated music and fraudulent streaming, making Tuesday’s sentence an early test of how the criminal-justice system will respond when artificial intelligence is used to industrialize music-streaming manipulation.

And the scale was extraordinary.

Smith wasn’t accused of buying a few thousand streams to make a song look popular. Prosecutors said he created hundreds of thousands of tracks, operated thousands of fake listener accounts and generated billions of fraudulent streams over a scheme that stretched from 2017 until 2024.

At certain points, according to the government, Smith had as many as 10,000 bot accounts operating simultaneously.

The Music Was Fake. The Royalty Checks Were Real.

Smith’s operation exploited one of the fundamental features of the streaming economy.

Spotify, Apple Music, Amazon Music, YouTube Music and other services distribute enormous amounts of money to music rights holders based in part on listening activity. A stream may be worth only a fraction of a cent, but those fractions become meaningful when multiplied by millions — or billions.

Smith found a way to manufacture that scale.

According to federal prosecutors, he created fake streaming accounts using fraudulent email addresses and improperly obtained debit cards. Software then controlled those accounts and repeatedly streamed music Smith owned or controlled.

There was an obvious problem with simply playing the same handful of tracks over and over.

That kind of abnormal activity could trigger streaming platforms’ fraud-detection systems.

Smith needed volume not only in listeners, but in music.

He needed an enormous catalog.

That is where artificial intelligence entered the scheme.

He Needed “A TON of Songs Fast”

According to prosecutors, Smith understood that generating huge numbers of streams on a small number of tracks could attract attention.

Instead, he wanted to spread his artificial listening activity across an enormous catalog so that individual songs wouldn’t necessarily appear suspicious.

In an email cited by prosecutors, Smith explained that he needed “a TON of songs fast” to get around streaming platforms’ anti-fraud systems.

Artificial intelligence provided the scale.

Smith eventually obtained hundreds of thousands of AI-generated tracks that could be uploaded and streamed by his bot network.

The bots would then continuously play those songs, creating what appeared on the platforms to be legitimate consumer listening activity.

But the listeners weren’t real.

The popularity wasn’t real.

The royalty payments were.

Over the life of the scheme, prosecutors said Smith’s AI-generated catalog was streamed billions of times, producing more than $8 million in fraudulent royalty payments.

At One Point His Bots Were Outstreaming Taylor Swift

One comparison included by federal prosecutors demonstrates just how enormous Smith’s automated operation became.

In April 2023, Taylor Swift’s entire catalog generated approximately 9.3 million YouTube Music streams through family plans.

During that same month, Smith’s bot accounts generated approximately 80.9 million family-plan streams of his AI-generated music on YouTube Music.

In other words, using that specific category of YouTube Music streams, Smith’s automated network generated more than eight times the monthly plays of Swift’s entire catalog.

That didn’t mean actual listeners suddenly preferred Smith’s obscure AI catalog to one of the biggest artists in the world.

It demonstrated exactly the opposite.

The numbers had become detached from genuine human demand.

Thousands of Fake Accounts

The scheme required more than AI music.

Smith needed fake consumers.

Prosecutors said he used thousands of bot accounts across streaming platforms and automated them to continuously play his music.

At times, as many as 10,000 accounts were operating simultaneously.

The system was designed to imitate legitimate consumer activity at enormous scale while distributing the streams across enough tracks to reduce the chances that any one song would set off alarms.

An earlier email cited in the case showed Smith thinking explicitly about that problem.

“In order to not raise any issues with the powers that be we need a TON of content with small amounts of Streams,” he wrote in 2018, according to prosecutors’ sentencing materials reported by Music Business Worldwide.

That sentence essentially described the architecture of the operation.

Don’t make one fake hit.

Make an enormous number of songs and give each one enough fraudulent activity to collectively generate serious money without making any individual track look impossibly successful.

AI Made the Scheme Scalable

That is what makes the case particularly important for the modern music business.

Streaming fraud existed before generative AI.

Bot farms existed before generative AI.

Artificially inflated plays existed before generative AI.

What AI changed was the ability to produce an enormous supply of music cheaply and rapidly enough to feed an equally enormous network of automated listeners.

A traditional fraudster attempting the same strategy might need thousands of musicians, producers, recordings or licensed tracks.

Smith could use artificial intelligence to produce hundreds of thousands of tracks.

The music itself became another automated component of the system.

Machines could generate the songs.

Machines could listen to the songs.

The streaming platforms could count the plays.

And royalty systems could generate real money from activity in which genuine human listeners played little or no role.

The Scheme Ran for Years

Federal prosecutors say Smith’s operation ran between approximately 2017 and 2024.

The AI component became increasingly important as the scheme developed.

Smith was initially indicted in September 2024 on charges including wire fraud conspiracy, wire fraud and money laundering conspiracy. He ultimately pleaded guilty in March 2026 to one count of conspiracy to commit wire fraud.

The guilty plea significantly reduced his sentencing exposure compared with the original charges.

The conspiracy count carried a statutory maximum of five years in prison.

Prosecutors nevertheless wanted a considerably longer sentence than the one Smith ultimately received.

Prosecutors Wanted at Least 46 Months

Federal prosecutors asked Judge Koeltl to sentence Smith to at least 46 months in prison.

That was the low end of the federal sentencing-guidelines range of 46 to 57 months identified in the case.

The U.S. Probation Office recommended 24 months.

Smith’s attorneys asked the court for probation — meaning no prison sentence at all.

Koeltl landed between those positions, sentencing Smith to 18 months behind bars.

The result means Smith received less than half of the prison time prosecutors requested.

But the sentence is still significant because of the precedent surrounding the prosecution.

Both sides described the case as the first criminal prosecution of its kind in the United States.

For the music industry, that made Smith’s sentencing about more than one defendant.

It was an early indication that large-scale streaming manipulation can move beyond platform penalties, royalty withholding or civil disputes and become a federal criminal case.

Smith Has to Forfeit More Than $8 Million

The financial consequence is even larger than the prison sentence.

Judge Koeltl ordered Smith to forfeit $8,091,843.64.

That is the loss amount Smith and prosecutors agreed upon.

The number doesn’t necessarily represent the total amount of money that ever passed through Smith’s operation.

According to government sentencing materials reported by Music Business Worldwide, Smith received more than $14 million in total payments from royalty-paying entities. The approximately $8.09 million loss figure was calculated after deductions that included estimates for legitimate streams and money Smith effectively returned to royalty pools by paying for the streaming accounts his bots used.

Smith’s attorneys disputed the government’s characterization of the financial impact and argued that the $8 million-plus loss figure was far greater than the amount Smith personally pocketed.

They did not, however, dispute the agreed forfeiture amount itself.

Where Does the Money Actually Come From?

This is where the case becomes particularly important for artists and songwriters.

Streaming royalties aren’t generally created as an unlimited pool of money that expands every time someone presses play.

On major services, rights holders typically receive portions of royalty pools based on their share of eligible listening activity and other contractual factors.

That means fraudulent streams can distort how a fixed or revenue-linked pool is divided.

If fake songs controlled by one person accumulate millions of fraudulent streams, those streams can increase that person’s share of royalty distributions while reducing the relative share available to legitimate rights holders.

The Music Fights Fraud Alliance made that argument directly to the court, saying each fraudulent stream diluted the value of legitimate streams occurring during the same accounting period.

That is why prosecutors characterized Smith’s $8 million as money diverted away from genuine artists and rights holders rather than money simply created out of nowhere.

Smith’s Lawyers Challenged That Characterization

The defense presented a different view of the harm.

Smith’s attorneys argued that no individual recording artist or songwriter had been shown to suffer an appreciable identifiable loss because of Smith’s conduct.

They also emphasized that streaming manipulation remains widespread across the music business and argued that Smith was being subjected to a novel prosecution in an industry that had struggled with fake streams for years.

The defense also stressed an important distinction that can easily get lost in the headline:

Creating music with AI isn’t itself what sent Smith to prison.

The fraud did.

Smith’s attorneys pointed to public statements from Spotify indicating that music isn’t automatically disqualified from royalties simply because AI tools were involved in its creation.

That distinction is central to understanding the case.

He Wasn’t Sentenced Simply for Making AI Music

Smith’s conviction should not be interpreted as a federal court declaring AI-generated songs illegal.

AI-generated music is already being uploaded to streaming platforms in enormous quantities, and the legal questions surrounding copyright, training data, voice cloning and ownership remain separate and rapidly evolving.

Smith’s criminal conduct involved manipulating streams through fake accounts to fraudulently obtain royalty money.

The AI songs gave him enough content to make that manipulation scalable.

The bots supplied the fake listeners.

Together, those systems created fraudulent streaming activity.

That is what produced the criminal prosecution.

The Platforms Included Some of Music’s Biggest Services

According to reporting on the government’s case, Smith’s bot operation touched services including Spotify, Apple Music, Amazon Music and YouTube Music.

The indictment originally described streaming activity across major platforms without framing any of those companies as participants in Smith’s fraud.

They were among the systems being manipulated.

Spotify previously told Music Business Worldwide that its own fraud-prevention measures limited Smith’s earnings on Spotify to roughly $60,000 of the broader multimillion-dollar amount alleged in the case.

That detail is significant because it illustrates the cat-and-mouse game now occurring between platforms and streaming manipulators.

Platforms build detection systems.

Fraudsters design behavior intended to avoid those systems.

Platforms change their rules.

Fraudsters adjust.

AI dramatically increases the speed at which the content side of that equation can operate.

Spotify Has Already Changed Its Royalty Rules

Spotify has introduced additional requirements intended in part to prevent extremely low-engagement content from extracting small amounts of money at massive scale.

Since 2024, tracks generally need at least 1,000 streams during the previous 12 months to qualify for recorded-music royalty payments on Spotify, along with a minimum number of unique listeners that Spotify does not publicly disclose.

Those policies make a scheme like Smith’s more difficult.

They don’t necessarily make it impossible.

At the height of his operation, Smith was allegedly running as many as 10,000 bot accounts.

Scale can potentially overcome thresholds designed to filter out low-volume manipulation, which is why streaming companies increasingly rely on fraud detection rather than simple stream-count requirements alone.

The Music Industry Wanted the Court to Send a Message

Several music-industry organizations weighed in before Smith’s sentencing.

The Music Fights Fraud Alliance told the court that the sentence would become an important early signal about the legal consequences of streaming fraud.

Prosecutors made a similar argument.

They warned that a sentence viewed as too lenient could send the message that massive streaming manipulation does not produce meaningful consequences.

After the sentence was imposed, the Recording Industry Association of America welcomed the court’s recognition of the seriousness of the conduct.

The Music Fights Fraud Alliance similarly said Smith had undermined trust in the industry and argued that the broader fight against streaming fraud would require continued cooperation between music companies, law enforcement and policymakers.

AI Music Is Exploding Across Streaming Services

Smith’s case arrives at a moment when the amount of AI-generated music entering streaming platforms is exploding.

The problem is no longer theoretical.

Streaming services are increasingly dealing with huge volumes of automatically generated tracks, some legitimate experiments and others associated with spam, impersonation or fraud.

The Financial Times reported this week that Sony Music had sought the removal of more than 260,000 unauthorized fake tracks from streaming services by the end of September, including AI-generated material imitating major artists.

Those impersonation cases are different from Smith’s scheme, but they share a common technological pressure point.

AI has made audio extremely cheap to manufacture at enormous scale.

That changes the economics of abuse.

The Future “Artist” Might Be a Server Talking to Another Server

Smith’s operation demonstrates the extreme version of that problem.

Imagine a music ecosystem where a computer generates a song.

Another automated system uploads it.

Thousands of bot accounts listen to it.

Algorithms register those streams.

Royalty systems calculate payments.

Money is transferred.

At no point does meaningful human demand need to exist.

That is essentially the problem prosecutors say Smith exploited.

The technology wasn’t merely helping someone make music.

It was helping create the inventory needed for an automated financial machine.

It Also Raises Questions About Streaming Charts and Popularity

Smith’s scheme was primarily about royalties, not turning himself into a celebrity.

In fact, spreading the streams across huge numbers of tracks was specifically designed to avoid drawing attention.

But the case highlights a broader vulnerability in digital music.

Streaming numbers function as more than royalty inputs.

They are also treated as signals of popularity.

Artists, labels, managers, booking agents, playlist curators and fans routinely use streaming statistics as evidence of momentum.

If those numbers can be manipulated at scale, the consequences can extend beyond direct royalty theft.

Artificial streams can potentially influence perceptions about which artists are growing, which tracks deserve playlist placement and which acts appear commercially valuable.

That is why streaming fraud has become a credibility problem as much as a financial one.

Smith’s Scheme Was Designed to Look Boring

Perhaps the most sophisticated aspect of the operation was that Smith apparently didn’t want a gigantic viral song.

That could have exposed him.

His goal was the opposite.

Hundreds of thousands of songs.

Relatively small amounts of suspicious activity distributed among them.

Thousands of accounts.

Continuous streams.

Month after month.

Year after year.

The strategy was essentially to make the fraud too fragmented to notice.

His 2018 email about needing “a TON of content with small amounts of Streams” captures the concept almost perfectly.

AI eventually made generating that ocean of content much easier.

Federal Investigators Eventually Caught Up

The FBI investigated the operation, and prosecutors from the Southern District of New York’s Complex Frauds and Cybercrime Unit brought the case.

Smith was indicted in 2024.

He pleaded guilty in March 2026.

Seven months later, Judge Koeltl sentenced him to 18 months.

Once he completes that sentence, Smith will remain under federal supervision for another two years.

And the $8,091,843.64 forfeiture judgment will remain a substantial financial consequence.

The Sentence Is Shorter Than Prosecutors Wanted — But the Precedent Is Bigger

Eighteen months may strike some people in the music industry as relatively modest considering the size and duration of the scheme.

Prosecutors wanted at least 46 months.

Probation officials recommended 24.

Smith wanted no prison time.

The judge chose 18.

But the most consequential part of the case may not be the precise number of months Smith spends behind bars.

It is the existence of the prosecution itself.

Streaming manipulation has long existed in an uncomfortable space where platforms can remove songs, withhold royalties, terminate accounts or pursue civil remedies.

Smith’s case crossed into federal criminal enforcement.

The Government’s Message Is Clear

U.S. Attorney Jamie McDonald framed the case as an attack on both the streaming economy and legitimate creators.

“By flooding music streaming platforms with automated bots in the place of consumers, and fake songs in the place of creativity,” McDonald said, Smith diverted millions in royalty payments from genuine artists.

The government’s framing matters.

Prosecutors aren’t treating fraudulent streaming as harmless manipulation of an algorithm.

They’re treating large-scale schemes designed to extract royalty payments as financial fraud.

And after Tuesday’s sentencing, there is now a federal prison sentence attached to that position.

This Could Be a Blueprint for Future Cases

Smith may be the first major U.S. defendant prosecuted over an AI-assisted streaming scheme.

He is unlikely to be the last if similar operations continue.

The technology required to generate thousands of songs is becoming easier to access.

Automation tools are becoming more capable.

Fake accounts can be operated at industrial scale.

And streaming services collectively distribute billions of dollars in royalties every year.

That creates an obvious incentive for fraud.

The Smith case now gives prosecutors a roadmap for how such conduct can be charged when they can establish intentional deception, manipulated streaming activity and fraudulent royalty payments.

The Real Story Isn’t That AI Made $8 Million

AI didn’t independently decide to steal streaming royalties.

Smith used AI as one component of a deliberately constructed fraud operation.

That distinction will become increasingly important as stories involving artificial intelligence and music multiply.

The technology generated content.

Bots generated listening activity.

Smith controlled the system and collected the money.

He ultimately admitted that conduct by pleading guilty to conspiracy to commit wire fraud.

The criminal liability came from the fraud, not from the mere existence of AI-generated music.

A Warning for the Streaming Era

The numbers alone make the case remarkable.

Hundreds of thousands of AI-generated songs.

Thousands of fake accounts.

As many as 10,000 bots operating at once.

Billions of streams.

More than $8 million in fraudulent royalties.

One federal conviction.

Eighteen months in prison.

Two years of supervised release.

And $8,091,843.64 in forfeiture.

But the larger question now facing the music industry is how many similar operations exist that haven’t been uncovered.

Smith’s system demonstrated that streaming fraud doesn’t necessarily need a famous artist, a viral song or even real fans.

It needs content, accounts, automation and enough patience to turn fractions of pennies into millions of dollars.

Artificial intelligence made the content easier to produce.

Bots supplied the audience.

For years, the royalty system supplied the money.

On Tuesday, a federal judge supplied the consequence.