Musician and composer

United States. The statistics bureau projects that employment of musicians and singers will show little or no change from 2025 to 2035, from 170,400 to 170,900, with about 17,400 openings a year. It says interest in live music is expected to support demand, and that declining public attendance of classical performances and reduced participation in church activities may slow growth; it names no technology or AI effect. It counts jobs for one country.

A named person with standing publicly predicted something, on a date, in an attributable statement. It is recorded so that who said what, and when, stays checkable — and it never moves a task's assessment, because a prediction is not an observation. Its value arrives later: the record sits on the same page as the evidence about that occupation, so anyone reading the forecast reads the record of what happened next beside it. That is the reckoning; this site publishes no verdict on whether a forecast came true.

United States. The statistics bureau projects that employment of music directors and composers will grow 1 percent from 2025 to 2035, from 51,100 to 51,700, with about 4,200 openings a year, and says composers will be needed to write film scores and music for television and commercials; the constraints it names are funding for orchestras and other groups, not technology. It counts jobs for one country.

A named person with standing publicly predicted something, on a date, in an attributable statement. It is recorded so that who said what, and when, stays checkable — and it never moves a task's assessment, because a prediction is not an observation. Its value arrives later: the record sits on the same page as the evidence about that occupation, so anyone reading the forecast reads the record of what happened next beside it. That is the reckoning; this site publishes no verdict on whether a forecast came true.

A music streaming service. It says tracks detected as AI-generated peaked in June at more than 50% of total uploads of new music, that it now receives nearly 90,000 fully AI-generated tracks every day, that such music accounts for only 1–3% of streams, and that the primary purpose of uploading these tracks appears to be fraudulent; it will take down AI tracks used for streaming fraud. Detection uses Deezer's own tool, which it also licenses to others.

An employer has put it into production. Can move the baseline — weighted by scale and how similar the setting is.

A music streaming service. It says daily deliveries of AI-generated music averaged around 60,000 tracks in January 2026, roughly 39% of all music delivered daily, that only a small share of total streams (up to 3%) came from AI-generated tracks, and that up to 85% of those streams were fraudulent in 2025 depending on the month, against 8% for the whole catalogue; fraudulent streams are excluded from royalty payments. It is the service's own policy and detection, which it also sells.

Failure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.

A major record label's announcement. It says the deal settles previous litigation with the AI music company Suno; that when new licensed models launch in 2026 the current models will be deprecated and downloading audio will require a paid account; and that artists and songwriters will have full control over whether and how their names, images, likenesses, voices and compositions are used in new AI-generated music. It is a party describing its own deal.

Failure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.

A major record label's announcement. It says UMG and the AI music company Udio settled copyright infringement litigation; that a new platform launching in 2026 will be trained on authorized and licensed music; and that Udio's existing product will remain available during the transition with creations controlled within a walled garden. It is a party describing its own deal; UMG continued to sue another AI music company.

Failure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.

A blind, randomised listening study using songs from real-world use of a commercial generator. The authors report that when pairs are random, listeners cannot differentiate AI music from human-made songs and are no better than random guessing, while they can when pairs are highly similar, and that practical musical experience increases identification. After attention checks 308 participants remained, mostly Portuguese speakers; it is a workshop-track paper.

A demo, benchmark or paper shows the task can be done. Updates what the technology can do — not what employers will do.

A music streaming service's policy announcement. It says that in the past 12 months, a period marked by the explosion of generative AI tools, it removed over 75 million spammy tracks; that vocal impersonation is only allowed in music on Spotify when the impersonated artist has authorised the usage; and that it will support an industry standard for disclosing where AI played a role in a track. Spam includes more than AI tracks, and the service also builds its own AI features.

Failure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.

Global. The confederation of authors' societies says a study it commissioned from PMP Strategy projects that by 2028 generative AI music will account for approximately 20% of traditional music streaming platforms' revenues and around 60% of music libraries' revenues, and that music creators will see 24% of their revenues at risk of loss by 2028. Its director general says the study was commissioned to show the value copyright works bring to AI companies; it is a model built by an interested party, not a measurement.

A named person with standing publicly predicted something, on a date, in an attributable statement. It is recorded so that who said what, and when, stays checkable — and it never moves a task's assessment, because a prediction is not an observation. Its value arrives later: the record sits on the same page as the evidence about that occupation, so anyone reading the forecast reads the record of what happened next beside it. That is the reckoning; this site publishes no verdict on whether a forecast came true.

United States and Canada, film and television scoring under the musicians' union contract. The agreement states that the term musicians includes only humans; requires the producer to give the union written notice when a motion picture's score is created in whole or in part using a generative-AI music soundtrack; and pays each musician who recorded the music used as a prompt three hours at the electronic multitracking rate for each fifteen minutes of generated soundtrack used. It compensates and discloses; it does not forbid generative use, and it excludes traditional tools such as virtual instruments.

Failure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.

Germany and France. A survey of 15,073 members of the two collecting societies found that 35% had used AI technologies in their work, including 52% of those making advertising music and 48% of those making library music; the study's expert interviews conclude that a substitution of human-made background music is already taking place, and it estimates that up to 27% of music creators' revenues are at risk by 2028. The societies commissioned the study and campaign for remuneration and transparency; use is self-reported.

Measured, large-scale use of a tool for real work, where the decision to use it was the worker's rather than an employer's. It is more than a capability record — the work is real, not a demo — and less than a deployment record, because no employer put it into production, required it, or built a process around it. Weighted `cautious`: `automating` means the machine can do the task AND there are adoption signs, and this is an adoption sign — but usage can be experimental, and much of the measurement comes from a party with a stake, so one record is never enough and two independent ones are. Note who is counting. Vendor telemetry sees this directly and sells the tool, so such a record names that stake in its scope; a statistics agency asking firms whether their workers use AI in tasks sees the same channel with no stake at all, and that is the better source where it exists.

If you are starting out, expect generated music to compete hardest for background, stock and advertising work, and build on what generated tracks do not get: live audiences, commissioned scores, and a name people seek out.

Expect library and advertising briefs to shrink or pay less, and your rights to matter more: licensing deals now decide whether your recordings and compositions train AI, and on what terms.

Stay and strengthen

Stay, and build around live and commissioned work

Live music is expected to support demand, and union scoring contracts define musicians as humans.

Reshape the role

Manage your rights in the AI licensing market

Labels are signing licences with AI firms that give artists opt-in control over their voices and compositions.

Adjacent move

Move towards music for picture under union terms

The US film musicians' agreement requires notice and payment when generative AI is used on a score.

Not across the board. Generated tracks are flooding streaming services and compete hardest with background and library music, but they draw only a small share of listening. Live performance, commissioned scores under union contracts and artists people seek out are less exposed, and US projections expect musicians' and singers' employment to show little or no change from 2025 to 2035.

We do not answer that with a number of years. Watch which of your income streams generated music competes with — library and advertising first — and what your contracts and licences say about AI. Those tell you more than any date.

Often not. In a controlled study with real songs from a commercial generator, listeners given random pairs did no better than guessing; when the pairs were similar, and when listeners had musical experience, they did better.

It is being settled by lawsuits and licences. Major labels sued AI music generators; Warner and Universal then settled with licensing deals, and Warner's gives artists and songwriters control over whether their voices and compositions are used. Ask your label, publisher or collecting society what applies to you.