AI in music: what LANDR's 87% and 46% figures actually mean
LANDR surveyed 1,241 music makers about AI. Here is what 87% usage and 46% positive sentiment toward creative AI mean—and what they do not.
Two numbers from LANDR’s 2025 report on AI in music are easy to place next to each other and easy to misread:
- 87% of respondents used AI somewhere in their music workflow.
- 46% had a very or somewhat positive view of the idea of using AI for creative tasks.
The second figure does not measure trust. The survey did not ask whether 46% of musicians “trust AI to create,” and the first figure does not mean that 87% use AI to write or generate songs.
That distinction changes the story. The report shows broad use across many technical, creative, and promotional tasks, alongside much more cautious sentiment when AI touches songwriting, ideation, or generated parts.
The two headline figures, precisely
| Figure | What LANDR measured | What it does not establish |
|---|---|---|
| 87% | Respondents who selected at least one AI use across production, creative, or release/promotion questions | That 87% use generative AI, write songs with AI, or approve of every use |
| 46% | Respondents who selected “very positive” or “somewhat positive” about the idea of AI for creative tasks | Trust, willingness to delegate a song, satisfaction with a tool, or actual creative usage |
Usage and sentiment are different variables. Someone can use an AI feature because it is built into a tool while remaining ambivalent about it. Another person can feel positive about a possible use they have not tried. The report measures both areas, but it does not prove why an individual answered as they did.
How LANDR ran the survey
The LANDR report is based on an online survey of 1,241 music makers from LANDR’s global community, aged 16 or older, conducted from September 30 to October 6, 2025. It included more than 30 multiple-choice, single-choice, and open-ended questions.
The sample covered different experience levels: 43% described themselves as advanced and 11% as expert. It was also weighted heavily toward some groups and regions—for example, 41% of respondents were in the United States, 22% in the UK, Canada, or Australia, and 24% in Europe.
That makes the report useful as a picture of LANDR’s surveyed community. It should not be treated as a precise estimate for every musician worldwide. The report does not present the sample as a probability sample of all music creators, so its percentages are best read with the population and wording of each question attached.
What the 87% usage figure includes
LANDR calculated the 87% figure from respondents who selected at least one use across a broad set of tasks. The report groups current use into three areas:
- 79% used AI for technical tasks.
- 66% used AI for creative tasks.
- 52% used AI for promotion.
These categories are broad. Technical tasks include functions such as restoration, timing correction, mixing/mastering, or stem separation. Creative questions cover songwriting, ideation, generated or edited parts, and related work. Promotion includes areas such as content ideas, bios, visuals, audience research, or release planning.
So the strongest defensible reading is: most people in this LANDR sample used at least one AI-enabled tool or feature somewhere in their workflow. It is not evidence that most of them hand complete songs to a generator.
What the 46% figure really says
For perception, LANDR asked how respondents felt about the idea of using AI in three areas. The percentage shown is the share choosing “very positive” or “somewhat positive” on a five-point scale:
- 77% for technical tasks.
- 70% for promotion and marketing.
- 46% for creative tasks.
The gap is meaningful, but it needs the right label. It shows less positive sentiment toward creative AI than toward technical or promotional AI. It does not tell us that the other 54% distrust AI, reject it, or would never use it. That remainder also includes neutral and different degrees of negative response, which the headline figure does not separate.
The report also does not establish the cause of the gap. Concerns about quality, ethics, dependency, platform rules, or human replacement may help explain the wider context, but the survey results presented do not prove that any single concern caused an individual respondent’s rating.
“Creative use” is not one behaviour
Among the creative tasks displayed in the report, current use was spread across many specific activities. The most selected included:
- creating lead vocals: 18% of all respondents;
- creating drum patterns: 16%;
- creating instrumental parts: 16%;
- extending or arranging ideas into a complete song: 14%;
- creating variations on an existing melody: 13%;
- creating song structures: 12%;
- generating melodies or toplines: 12%;
- generating chord progressions: 12%.
These were multiple-choice questions, so the percentages are not mutually exclusive and should not be added together. LANDR also notes that the slide displays only the most selected options among a longer list.
This matters because “uses AI creatively” can describe very different workflows: polishing a lyric, generating a drum pattern, exploring a structure, or using a generated vocal. A single headline cannot tell us how much creative control a person keeps or how central AI is to the finished work.
Song generators: current use and openness are also different
LANDR reports that 29% of respondents already used song generators such as Suno or Udio at some stage of their workflow. A further 12% were very open to trying one soon and 24% said they might try one but were unsure. Together, that produces the report’s 65% “using or open to using” figure.
Again, those groups should not be collapsed into one claim about adoption. “Already use,” “very open,” and “might try” describe different levels of behaviour and intent.
When the report asked which parts people used from song generators, the displayed responses included vocals (16%), instruments (14%), song structure or arrangement (14%), beats or instrumentals (13%), and an entire track (13%). These figures use all respondents as the base. They show that full-track use exists, but they do not support the claim that most generator users rely only on isolated parts unless the data are re-analysed specifically among users.
Benefits and concerns reported by respondents
LANDR asked people to choose their biggest benefits and concerns from multiple-choice lists. Respondents could select top options, so these percentages overlap.
Most selected benefits
| Benefit | Share of all respondents |
|---|---|
| Fill skill gaps | 38% |
| Work faster | 33% |
| Automate tasks they dislike | 29% |
| Get inspiration | 28% |
| Learn new techniques or genres | 23% |
Most selected concerns
| Concern | Share of all respondents |
|---|---|
| Generic, soulless, or low-quality music | 46% |
| Ethics, including use of artists’ work without consent | 43% |
| Dependence on technology | 34% |
| Release rules or platform takedowns | 30% |
| AI replacing humans | 29% |
It is fair to say that quality and ethics were the two most selected concerns shown. It is not fair to infer from these responses that every respondent who disliked creative AI did so for those reasons, or that the survey settled those debates.
Is usage likely to keep growing within this sample?
The report found that 69% of respondents agreed that they were using more AI tools than a year earlier. Within that group, 90% expected to use more AI tools over the next 12 months. Among the 31% who had not increased their use, 24% expected to use more.
That indicates very different trajectories between respondents already deepening their use and those who are not. It is evidence of reported intention, not a forecast guaranteed to occur. It also comes from the same LANDR community sample, so it should not be presented as a universal adoption curve for the whole music industry.
What the report supports—and what it does not
Supported by the published survey
- AI-enabled tools were widely used somewhere in the surveyed workflows.
- Current use was higher for technical tasks than for creative or promotional tasks.
- Positive sentiment was lower for creative uses than for technical or promotional uses.
- Respondents identified both practical benefits and serious concerns.
- People already increasing their AI use were more likely to say they would increase it again.
Not established by the survey
- That 46% of musicians trust AI to create music.
- That 54% distrust or reject creative AI.
- That 87% use AI to write songs.
- That technical AI is “mandatory” for independent musicians.
- That AI users will make better music or gain a competitive advantage.
- That musicians prefer assistance over generation as a universal rule.
- That the data validate any specific company’s product or mission.
Zoundroom’s interpretation: assistance should remain a choice
The following is Zoundroom’s product perspective, not a finding from LANDR’s survey.
We think creative AI is most useful when the musician can decide when to ask for help, inspect the suggestion, and reject it without losing the thread of the song. That is why Zoundroom’s assistant is positioned as a copilot inside a project rather than as a system that composes a finished track for you.
According to Zoundroom’s current product page, the assistant can propose chord progressions, lyric ideas, song structures, rhymes, synonyms, and feedback on how lyrics may be perceived. The musician decides what to keep. Those are product claims about Zoundroom; LANDR did not evaluate the app and the survey does not prove that this design is preferred by all musicians.
The more careful conclusion is not “musicians trust assistants but reject generators.” It is that the report captures broad adoption, uneven sentiment, and many distinct ways of using AI. Product teams and musicians still have to make deliberate choices about where the tool belongs and where it does not.
The bottom line
LANDR’s 87% and 46% figures are not contradictory once their questions are kept separate. One measures whether respondents used at least one AI tool or feature somewhere in a broad workflow. The other measures positive sentiment toward the idea of AI for creative tasks.
That is a more nuanced—and more useful—picture than a claim about trust. AI was already present in most surveyed workflows, but creative uses remained the area where positive sentiment was weakest and questions about quality, ethics, control, and dependency were most relevant.