Major Labels Push to Keep Fully AI-Generated Songs Off Global Music Charts

The Engadget major labels report says record companies want fully AI-generated songs barred from official charts, despite their growing presence on streaming platforms.

That demand turns a technical classification problem into a fight over cultural legitimacy. A song can attract streams, appear popular, and generate revenue without meeting an industry’s definition of human creativity.

The immediate conflict pits record labels and artist groups against systems that measure popularity without asking how a recording was created. Sweden has already excluded a predominantly AI-generated hit, while other major chart organizations have not adopted equivalent rules.

The Engadget Major Labels Story Is About Chart Eligibility

The proposed boundary would separate access to streaming services from eligibility for official music charts.

Streaming platforms generally accept recordings from distributors, count valid plays, and surface popular material through playlists or rankings. Official charts then use sales, downloads, or eligible streams to measure a market’s most popular recordings.

These systems overlap, but they are not identical. A platform ranking reflects activity inside one service, while an official chart applies its own eligibility and verification rules.

That distinction became visible in Sweden during January 2026. “Jag Vet, Du Är Inte Min,” credited to the project Jacub, reached the top of Spotify’s Swedish ranking after accumulating millions of streams.

Investigative reporting connected the project to people working with the Danish music company Stellar Music. Its creators acknowledged using AI while arguing that experienced human songwriters and producers directed the creative process.

IFPI Sweden nevertheless excluded the recording from Sverigetopplistan, the country’s official chart. Chief executive Ludvig Werner said a song that is mainly AI-generated does not qualify for the list.

The chart exclusion established an important precedent. Commercial listening activity remained real, but the official chart refused to recognize that activity under its existing creative criteria.

The controversy now extends beyond one Swedish recording. Industry groups want a consistent way to distinguish music generated primarily through prompts from music created by people using AI-assisted production tools.

On July 10, 2026, IFPI, RIAA, A2IM, WIN, IMPALA, the Recording Academy, SAG-AFTRA, and the Human Artistry Campaign announced a common labeling approach. It defines separate “AI-Generated” and “AI-Assisted” categories.

An AI-generated label applies when the core sound recording comes from generative systems without meaningful human performance. AI-assisted describes recordings where human creativity remains central, although generative tools contribute to production.

The labeling framework is voluntary and designed for adoption across streaming services and other music partners. It does not create a binding worldwide chart ban.

That limitation matters. Labels can supply standardized metadata, but individual chart organizations still decide which recordings qualify and how supporting evidence gets verified.

The Engadget major labels story therefore concerns a policy campaign, not a completed global prohibition. The industry has created a vocabulary for classification, while enforcement remains fragmented across markets.

The real change is that AI provenance, meaning the documented origin of a recording, is becoming a formal eligibility question. Popularity alone no longer settles whether a track deserves official recognition.

The Upload Flood Is Forcing the Decision

AI music has become an infrastructure problem because generating recordings is much faster than reviewing, labeling, or promoting them.

Deezer said fully AI-generated music represented more than half of its daily uploads by June 2026. The company measured a monthly average of roughly 90,000 such tracks arriving each day.

That was a sharp increase from January 2025. Deezer then identified approximately 10,000 fully generated tracks each day, representing about 10 percent of incoming music.

By April 2026, the company said the figure had reached nearly 75,000 tracks, or 44 percent of daily uploads. The latest upload count suggests synthetic production continued growing faster than the wider catalog.

These figures describe Deezer’s intake, not every streaming platform. They also depend on the company’s detection technology and definition of a fully generated recording.

Even with those qualifications, the direction is clear. Music generation tools can produce more recordings than traditional distribution and review systems were designed to handle.

Human musicians face practical limits involving writing, rehearsal, recording, mixing, and release schedules. Generative systems can create multiple complete tracks from text instructions during the same period.

Distribution services add another force multiplier. A creator can send one recording to several platforms without working through a record label or a physical manufacturing process.

The resulting volume does not prove that listeners want AI music. Research published in June 2026 examined AI-associated tracks on Spotify and found that 93 percent received few or no plays.

That finding challenges the assumption that upload volume equals consumer demand. Most generated songs disappear into the catalog, just as many human recordings attract limited audiences.

However, low engagement across most tracks does not eliminate the systemic risk. Large catalogs give operators more opportunities to find a viral sound, exploit recommendation systems, or combine synthetic music with artificial streams.

The economics favor scale. A person does not need every song to succeed when software can generate and distribute thousands of candidates.

This asymmetry pressures streaming services first. They must store files, process metadata, review rights claims, detect impersonation, evaluate fraud, and respond to takedown requests.

Chart organizations face a narrower but more visible problem. A single synthetic hit can challenge the public meaning of a ranking that claims to represent popular music.

The labels also have a direct commercial interest. They spend money developing artists, recording music, marketing releases, and securing attention inside crowded services.

An unlimited supply of low-cost tracks increases competition for that attention. It can also dilute royalty pools when payment systems distribute fixed subscription revenue across eligible streams.

None of this establishes that every AI-generated song is fraudulent. A creator can use a generator, disclose that use, attract genuine listeners, and avoid impersonating anyone.

The pressure comes from the combination of scale, weak disclosure, and uncertain authorship. Existing systems often cannot distinguish that legitimate case from coordinated catalog spam.

Major labels want chart rules to create a visible boundary while platforms develop stronger controls. A chart exclusion is simpler than proving how every sound was created or how every listener discovered it.

Human Creativity and Measured Popularity Are the Main Opponents

The central dispute is whether charts should record audience behavior or enforce a definition of authorship.

A measurement-first position starts with a simple argument. If real people voluntarily stream or purchase a track, a chart should reflect that activity regardless of the production method.

Under this view, excluding a popular recording makes the chart less accurate. It replaces observable consumer behavior with an institution’s judgment about acceptable creative tools.

Charts already include music shaped by digital production. Drum machines, sampling, pitch correction, software instruments, and algorithmic mastering have changed how recordings are made.

AI can also assist a human-led project without replacing its creators. A musician might use it to clean audio, suggest arrangements, generate a supporting texture, or restore an old performance.

The Beatles’ “Now and Then” illustrated the distinction in 2023. Machine learning helped isolate John Lennon’s voice from an old demo, but people still wrote, performed, and completed the song.

A human-centered position draws the boundary elsewhere. It argues that a recording generated mainly from prompts is not equivalent to one built through human performance and artistic decisions.

Industry groups say fans deserve to know whether generative AI produced the music they hear. Their new labels attempt to preserve this distinction without treating every use of AI alike.

That separation also supports existing Grammy rules. The Recording Academy allows AI-assisted work but requires meaningful human authorship for award eligibility.

Official charts are not awards, though. They usually measure market activity rather than artistic merit, which makes the case for exclusion more complicated.

The UK’s Official Charts Company states that releases must follow eligibility rules intended to protect chart integrity and reflect genuine sales. Those rules aim to reflect genuine sales and consumption, not every activity recorded by retailers.

This means charts already filter raw behavior. They can reject manipulated transactions, impose format requirements, and determine which streams count toward a published ranking.

The question is whether production method belongs beside those controls. Fraud describes manipulated consumption, while AI generation describes how a recording was made.

Combining the two would create an unfair presumption. A disclosed synthetic song with genuine fans is not the same as a human recording promoted through bot streams.

Separating them entirely creates another problem. Generative systems make it easier to produce catalogs designed around statistical experimentation rather than sustained artistic identity.

A chart exclusion therefore expresses a cultural choice. It says official rankings represent popular human-centered recordings, not simply the most consumed audio files.

That choice benefits established labels, but it also benefits independent human musicians competing for limited visibility. Both groups face the same flood of automated releases.

Conversely, the rule would disadvantage creators who use AI because they lack studios, instruments, collaborators, or conventional performance skills. Some of them will view generation as a new form of composition.

The Engadget major labels framing captures the political imbalance. Record companies possess the institutional access needed to influence chart rules, while independent AI creators lack comparable representation.

Still, the industry’s preferred distinction is not simply labels against outsiders. Major labels have pursued licensed AI products and partnerships while opposing unlicensed training, impersonation, and undisclosed generation.

Sony, Warner, and Universal have all explored agreements that permit controlled AI uses. Those moves show that the industry is contesting ownership and governance, not rejecting every synthetic tool.

The main opponent is therefore not humans against machines. It is human-centered eligibility against popularity measured without creative context.

AI-Generated Songs Are Not Automatically Streaming Fraud

Banning generated tracks from charts would address visibility, but it would not solve the economic abuse surrounding synthetic catalogs.

Streaming fraud occurs when operators manipulate listening activity to obtain royalties or influence rankings. It can involve bot accounts, stolen identities, artificial playlists, or repeated automated plays.

Generative music lowers the cost of producing material for those schemes. It does not create the fraudulent listening activity by itself.

A major United States case showed how the two can work together. Federal prosecutors accused Michael Smith of using hundreds of thousands of AI-generated songs and automated accounts to collect royalties.

Smith later pleaded guilty to conspiracy to commit wire fraud. According to the U.S. Attorney’s Office for the Southern District of New York, the operation generated more than $8 million in royalties.

The case did not turn on whether synthetic music could qualify as art. It focused on manipulated streams and deceptive royalty claims.

That distinction should guide platform enforcement. Services need to examine consumption patterns, account networks, ownership claims, and payment behavior alongside content detection.

Deezer says it tags fully AI-generated tracks and removes them from algorithmic recommendations. It has also said most streams involving detected AI music are fraudulent.

Tidal has taken a different economic approach. In June 2026, it announced labels for wholly generated music and said qualifying tracks would not receive royalties.

These policies go further than disclosure. They change how synthetic recordings are discovered and compensated, even when individual tracks have not been tied to fraudulent streams.

Spotify has emphasized conduct rather than production method. Its representatives have said AI use is not inherently the problem, while impersonation, unauthorized cloning, and fraud violate its policies.

Each strategy creates tradeoffs. Content-based restrictions are easier to explain, but they can penalize legitimate creators because of the tools they use.

Behavior-based enforcement is more targeted. However, sophisticated operators can distribute suspicious activity across many tracks and accounts to avoid obvious detection thresholds.

Chart exclusion sits between these approaches. It removes one public reward without necessarily deleting a recording or stopping royalty payments.

That could reduce incentives for synthetic projects designed to manufacture a visible hit. It would do less against operations pursuing many small royalty payments below chart level.

The policy might even obscure genuine popularity. An excluded track could dominate streaming activity while disappearing from the official chart that journalists and broadcasters use.

Separate AI charts are one proposed answer. They would preserve a public record of demand while keeping fully generated material outside human-centered rankings.

However, separate charts could become promotional targets for the same high-volume operators. Their value would depend on strong anti-fraud controls and reliable disclosure.

A universal ban would also face classification disputes. A song may combine human lyrics, generated vocals, live instruments, algorithmic mastering, and extensive human editing.

Detection software cannot always reconstruct that workflow from the final audio. Metadata and creator declarations will remain necessary, especially for mixed production.

Those declarations create incentives to underreport AI involvement. Distributors and chart bodies will need documentation requirements, audits, and penalties for false submissions.

The new labels help because they establish shared terms. They do not independently verify that creators have selected the correct category.

This is the largest weakness in the Engadget major labels proposal. The policy sounds clear when comparing a live band with a one-prompt song, but real production exists between those extremes.

A Global Ban Would Be Hard to Define and Enforce

The industry can agree on a principle faster than hundreds of platforms, distributors, and chart organizations can implement it.

Music charts operate under different ownership structures and market rules. Some are managed by national industry groups, while commercial publications or measurement companies run others.

A worldwide exclusion would require each organization to amend its eligibility standards. IFPI can coordinate national groups, but it cannot impose one rule on every independent chart.

The United States presents an especially important test. Billboard has included AI-generated recordings in sales and streaming charts when their measured activity met existing requirements.

An AI country act called Breaking Rust reached the top of a Billboard digital sales chart during 2025. Such results intensified questions about whether those rankings should distinguish synthetic performers.

Chart eligibility also depends on the underlying metric. Digital sales, on-demand streams, radio play, and social engagement represent different user actions and manipulation risks.

A platform can identify suspicious streams without determining whether a song has meaningful human authorship. A chart organization attempting both tasks inherits a much larger verification burden.

The industry’s “AI-Generated” and “AI-Assisted” categories provide a starting point. Yet phrases such as meaningful human input still require interpretation.

Consider a songwriter who writes every lyric and melody, then uses a generator for vocals and instrumentation. The concept is human-led, while the recorded performance is synthetic.

Now consider a producer who generates an entire track, replaces its vocal, rearranges every section, and remixes the result. The source is synthetic, but the finished recording reflects substantial human work.

Neither example fits comfortably into a binary system. Rules based on lead vocals or principal instrumentation will still produce contested edge cases.

Voice cloning adds another layer. A recording can involve extensive human direction while unlawfully imitating a recognizable performer.

That case raises rights and identity questions independent of chart eligibility. Excluding the song can limit exposure, but rightsholders may still need takedowns or legal action.

Detection accuracy is another uncertainty. Deezer has invested in identifying outputs from prominent generation models, but tools change frequently and can be modified after generation.

Human post-production can conceal synthetic signals. New models may also produce recordings that fall outside a detector’s training data.

False positives carry serious consequences. A mistaken classification could remove a human artist from recommendations, royalties, or chart consideration.

False negatives would undermine the rule by allowing undisclosed synthetic material to compete. Bad actors would have a reason to test detectors and alter their outputs.

A workable policy therefore needs an appeal process. Artists should be able to provide project files, performance recordings, session documentation, or other evidence supporting their classification.

Chart bodies will also need consistent consequences. A simple correction may be appropriate for a metadata error, while deliberate misrepresentation should trigger stronger penalties.

Public transparency matters just as much. Fans need to know whether a track was excluded for AI generation, fraudulent consumption, rights infringement, or missing documentation.

Without that detail, “AI slop” becomes a catch-all label. It can collapse artistic criticism, legal disputes, production techniques, and financial misconduct into one emotionally loaded category.

The term describes low-value material produced at scale, but it is not a technical standard. Rules should focus on observable production and conduct rather than subjective judgments about quality.

The strongest policy would combine standardized disclosures with behavioral fraud detection. It would then let each chart clearly state whether it represents all recordings or human-centered releases.

That approach would not satisfy everyone. It would at least make the boundary visible, reviewable, and less dependent on guesswork.

What the Fight Means for Artists, Platforms, and Listeners

The outcome will determine whether creative provenance becomes as important as sales and streams in music’s public rankings.

Human artists stand to gain from stronger disclosure. Labels can help listeners distinguish a performed recording from one produced mainly through generation.

Independent musicians may benefit even more because they lack large marketing budgets. Reducing automated catalog spam can improve their chances of reaching curators and real audiences.

However, independent creators also face compliance costs. A large label can maintain detailed production records, while a bedroom producer may struggle with new metadata and evidence requirements.

AI creators face the clearest restriction. They could retain access to streaming platforms yet lose entry to influential rankings, awards, editorial playlists, or royalty programs.

That would create a two-tier market. Synthetic recordings could remain available, but institutional systems would treat them as a separate category.

Streaming services must decide whether neutral hosting remains sustainable. The rapid growth of generated uploads increases storage, review, detection, and rights-management work.

Platforms also risk losing listener trust. A consumer who believes a playlist features emerging human artists may react negatively upon discovering undisclosed fictional performers.

Clear labels offer users a choice without requiring a complete ban. Services could let listeners include, exclude, or separately browse generated recordings.

Those controls would produce useful demand data. The industry could learn whether consumers reject AI music itself or mainly object to deception.

Listeners have a stake in preserving accurate charts. Rankings influence radio programming, media coverage, licensing decisions, touring opportunities, and perceptions of cultural momentum.

At the same time, listeners may reasonably expect charts to reflect what people actually consume. Removing eligible streams for policy reasons changes the meaning of those rankings.

The best solution may involve parallel reporting. A chart could publish its human-centered ranking while separately disclosing excluded synthetic activity.

That would preserve transparency without granting identical status. It would also prevent an officially invisible hit from distorting public understanding of listener behavior.

The debate has implications beyond music. Publishers, video platforms, app stores, and search engines face similar questions about ranking abundant machine-generated material.

Each system must decide whether it measures attention alone or incorporates provenance, effort, rights, and authenticity. Those choices shape which kinds of production remain economically viable.

For knowledge workers, the lesson is familiar. When creation becomes cheap, verification and organization become more valuable.

The scarcity moves from content to trusted context. Audiences need reliable records showing where material came from, how it was produced, and whether its engagement is authentic.

Music charts are among the first cultural institutions forced to make that decision publicly. Their answer will influence how other creative markets treat machine-generated abundance.

Three Signals Will Show Whether the Ban Becomes Real

The next phase depends on chart rules, reliable metadata, and evidence that exclusions change incentives rather than appearances.

The first signal is whether Billboard, the UK Official Charts Company, and additional IFPI-affiliated charts publish explicit AI eligibility standards.

Sweden has already demonstrated that exclusion is possible. A scattered collection of national decisions will not create the worldwide boundary sought by major labels.

Formal rules must define the threshold between generated and assisted music. They must also explain how creators can prove human performance and appeal disputed classifications.

The second signal is adoption of standardized AI metadata by distributors and major streaming services. Voluntary labels matter only when they travel with a recording across the supply chain.

Apple Music, Deezer, Tidal, Spotify, and digital distributors do not currently apply identical policies. Their choices determine whether chart bodies receive consistent information.

Watch for mandatory declarations during upload, visible listener labels, and penalties for inaccurate disclosures. Those changes would strengthen the industry’s position.

The third signal is whether fraud and catalog volume decline after new restrictions. A successful policy should produce measurable changes beyond keeping synthetic songs off public rankings.

Deezer’s daily upload totals provide one useful benchmark. Suspicious-stream rates, disputed royalty payments, and impersonation takedowns would offer additional evidence.

If generated uploads continue rising while chart appearances fall, the policy will have changed visibility without changing the underlying economics.

If fraudulent activity also declines, chart exclusion may be reducing incentives within a broader enforcement system. That would support extending the approach across more markets.

If legitimate AI creators attract real audiences outside official charts, pressure will grow for separate rankings or revised eligibility rules.

The Engadget major labels report captures the opening move in this contest, not its resolution. Labels have defined the category they want institutions to exclude.

They have not yet secured a universal rule, a perfect detection system, or agreement about recordings that combine human and synthetic elements.

The decisive question is no longer whether AI-generated music exists. Streaming platforms now receive tens of thousands of such recordings every day.

The question is what an official chart promises to measure. Is it a record of every genuine listening choice, or a ranking reserved for human-centered creative work?

Chart organizations should answer that question directly, publish their evidence standards, and show excluded activity rather than hiding it. Listeners can then judge both the music and the rules shaping what becomes a hit.