Ditto survey: 74% of over-55 independent artists use AI, three times the rate of 18–24s

Ditto Music published its Independent Music Report 2026 on 29 Sep 2026, built on answers from 5,039 self-releasing artists in 112 countries. Among the findings on money and touring was one that runs against the usual picture of AI music as a young person's prompt game: AI use rises with age. The report says 74% of over-55s use AI in their music-making, against 24% of 18–24-year-olds. Ditto's press release puts it as over-55s, "not Gen Z", driving adoption.

We covered the same report's income findings separately. This piece is about who uses AI, and for what.

What did Ditto find?

According to the report:

• 47% of respondents use AI somewhere in their music-making.
• Over-55s are more than three times as likely as 18–24s to say they use it, and the report says the same age divide shows up among US respondents alone.
• Production tasks such as mixing, mastering and vocal tuning are the most common use. Cover art comes second.
• Almost 1 in 4 artists who use AI have made fully AI-generated music, or 11% of the whole sample.

Ditto links its raw, respondent-level data from the report, so we downloaded it and counted for ourselves. The headline numbers match: by our count, 346 of 470 over-55s and 261 of 1,096 18–24-year-olds said they use AI tools. The climb is steady in between, at roughly 40% of 25–34s, 60% of 35–44s and 69% of 45–54s. The gap is wider for regular use: about 41% of over-55s use AI tools regularly, against about 10% of 18–24s.

How much weight can these numbers carry?

Less than the headline suggests, for four reasons.

The sample is Ditto's customers. The report's methodology says everyone surveyed had released music through Ditto and was recruited through its channels, and that the results should be read as a survey of active, self-releasing artists. It gives no fieldwork dates.

The over-55 group is small. It is 470 people, under a tenth of the sample. "Led by over-55s" describes a rate, not a crowd: in our count there are about twice as many AI users aged 25–34 (679) as over 55.

The "mixing and mastering" framing is a reading of a broader answer. In the raw data, the option most AI users ticked is "Composition & production (mixing, mastering, vocal tuning)". Composition and mixing share a box, so the survey can't tell an artist who has AI master a finished song from one who has it write the chords. Older AI users weren't clearly more cautious, either. By our count, about 21% of over-55 AI users ticked "I make fully-AI generated music", close to the 19% of 18–24 users who did.

Ditto sells AI tools. Its site offers an AI mastering service at $5 to download each mastered track. That doesn't make the numbers wrong, but a distributor with an AI mastering product has a reason to present AI as a helpful production tool.

The general shape does match other surveys of producers. Tracklib's 2025 study of 1,734 producers, as reported by MusicTech, found 32% using AI tools, mostly for jobs such as stem separation, EQ and mastering. Only 6% used fully generative tools, and 82% opposed generating whole songs from text prompts. We couldn't open Tracklib's own release (the link 404s) and the coverage gives no age breakdown. Ditto's own August 2025 poll of more than 2,000 artists and music professionals points the same way: only 10% said "definitely yes" to listening to fully AI-generated tracks, and 31% said they would avoid AI music even from a favourite artist.

What artists are saying

The argument among artists is less about age than about where a tool ends and authorship begins. One X user on X mocked producers who build tracks from sample packs, Splice loops and presets but object to AI music, saying almost nothing in their setup is their own creation. Another drew the line at control: the more settings an artist adjusts, the more intent shows and the more it counts as art. Splice's sample variations were acceptable to them; Suno was not.

We found no first-hand accounts from over-55 artists explaining why they use AI. The age story rests on Ditto's numbers alone.

What we think

The most interesting line in Ditto's data isn't in its report. When we split AI use by how long people have been making and releasing music, the artists most likely to use AI are the newest ones. About 62% of respondents releasing for less than a year use AI, against 35–50% in every longer-serving group. Among over-55s, 219 of 470 (47%) have been releasing for under three years, and 101 of the 111 over-55s releasing for less than a year said they use AI. These subgroups are small and self-selected, so treat this as a lead, not a finding.

Our reading, which is inference and not something Ditto tested: a large share of the "older AI adopters" are late starters. They write and perform songs, but they never had a mix engineer or a label budget behind them, and never spent their twenties learning to mix in a DAW. With 74% of the whole sample earning under $1,000 from music last year, according to the report, paying a person to mix and master every single isn't realistic. AI fills the engineer's chair they never had. It isn't writing their songs for them.

If that's right, the people AI music tools matter most to are not the ones the generative-music debate is about. They need three things the market doesn't reliably give them:

1. Tools that explain themselves. A newcomer can't judge a master they don't understand. Tools that show each EQ and compression move, and let you undo it, also teach you to mix.
2. No rights grabs. Before you upload stems to any AI mixing or mastering service, read its terms for licences over your audio, prompts or settings, and keep your unprocessed files.
3. Assistance by default, not generation. A late starter's songs are the point. Tools should finish the engineering and leave the writing to them.

Veena is in this space too. It offers AI-assisted mixing and real-time vocal tuning, but its CoProducer also generates parts layer by layer, so it sits on both sides of the line we're drawing. It should be judged on the same three tests.

For Ditto: publish the age-by-use-case breakdown and separate composition from mixing in next year's question. Until then, "over-55s use AI to mix" is a plausible story the survey can't fully support.