American producer Michael Smith is seeking to avoid prison after pleading guilty in a streaming-fraud case involving bots, AI-generated music and more than $10 million in royalties, according to Music Ally. Smith pleaded guilty in March to one count of conspiracy to commit wire fraud and agreed to forfeit more than $8 million as part of his plea.
His legal team has argued that the fraud did not cause widespread harm to songwriters and artists because any individual fraudulent stream has only a tiny effect on another artist’s share of total streams. Whatever the court ultimately decides at sentencing, the reported scale of the operation exposes a much larger problem for the music business.
When music can be generated at scale and listening can also be automated at scale, a system can create both sides of what looks like a market. The music exists. The streams exist. Money moves. But the audience those numbers appear to represent may not.
A stream is not the same as a customer
Streaming systems use listening activity as one of the signals through which economic value is calculated. That works only as well as the relationship between the activity being measured and actual audience behavior.
Bots weaken that relationship. A single artificial stream may be economically insignificant, but millions of automated streams can transform manufactured activity into a meaningful royalty claim. The number gets bigger without a corresponding increase in people actually choosing to listen.
AI can make the problem more scalable. Music can be produced or assisted by automated systems in enormous quantities, while bots can manufacture apparent demand for that material. The result is a feedback loop in which automated supply meets automated attention and produces real economic consequences.
That makes an increasingly important distinction visible: attention is not the same thing as a customer relationship.
The problem is bigger than fake streams
Streaming fraud is usually discussed as a detection problem. Platforms need to identify suspicious activity, remove fraudulent streams and prevent money from being diverted through manipulation.
Those protections matter, but the Smith case points to a deeper structural issue. Creators have built businesses inside systems where measurements produced by an intermediary can become some of the most important evidence of whether their work has an audience.
Streams, views, followers, likes and other engagement signals can all be useful. But creators generally do not operate the infrastructure that produces those numbers. The platform defines the activity, measures it, determines which activity qualifies and controls how that information affects discovery or compensation.
That means creators can become dependent not only on the platform for distribution, but also on the platform’s representation of their relationship with their audience.
When automated systems can manufacture some of those signals, that dependency becomes harder to ignore.
Direct support is a different kind of signal
A fan choosing to buy something from an artist is fundamentally different from another increment on a stream counter. It represents an economic decision rather than simply an instance of measured attention.
That does not mean every payment proves that a unique human being is genuine, nor does direct commerce make fraud impossible. Fraud can exist anywhere money moves. The distinction is that creators do not have to build their entire understanding of demand around engagement metrics generated and interpreted inside somebody else’s platform.
A direct purchase, paid access or tip creates an economic relationship between a supporter and a creator. For the creator, that can provide information that raw attention cannot: someone valued the work enough to act.
As automated content and automated engagement become easier to produce, that distinction becomes more valuable.
Creator-operated infrastructure changes the starting point
Certifyd approaches this problem from a different architectural direction.
Certifyd Core gives creators infrastructure they operate themselves. Identity, creative work and supported direct-commerce activity can begin from that creator-controlled foundation rather than requiring the creator’s entire business to exist as an account inside a centralized platform.
That does not make streaming unnecessary, and it does not make streaming fraud disappear. Fans can still discover and experience creative work through applications and services that provide valuable interfaces and audiences.
The difference is that those services do not have to become the permanent center of the creator’s business.
A creator selling directly to a supporter does not need millions of platform-generated interactions to demonstrate that an economic relationship occurred. The transaction is part of the creator’s own business activity rather than another engagement signal held entirely inside an intermediary’s system.
Certifyd is building toward extending that principle across a creator-operated network. Instead of every creator function accumulating inside one central platform, creators and independent providers can operate infrastructure and connect with one another while applications and services participate around them.
The goal is not to build a better stream counter. It is to give creators more ways to operate without making somebody else’s measurement of attention the foundation of their business.
When both sides can be automated
The Michael Smith streaming-fraud case is striking because it brings together two technological changes that are often discussed separately.
AI can dramatically increase the amount of content that can be produced. Automation can dramatically increase the amount of apparent engagement that content receives. Put them together and a system can begin manufacturing both the thing being consumed and evidence that somebody consumed it.
That does not mean streaming metrics have no value. Real people stream music every day, and streaming remains an important way audiences discover and listen to artists. It means the number itself cannot tell the entire story of the relationship between a creator and an audience.
For creators, that makes direct relationships increasingly important. A million streams can represent enormous genuine interest, manipulated activity or some mixture of the two. A creator should not have to build their entire business around figuring out which is which from numbers controlled by somebody else.
The future of the creator economy therefore cannot be measured only by how much attention a platform records. It also matters whether creators can establish relationships, conduct commerce and operate their businesses through infrastructure they control.
A stream can measure a listen.
It cannot, by itself, tell you whether you have a customer.