Study shows “78% of musicians are now using AI” – but is everything really as it seems?“78% of professional musicians now use AI.”
When that stat landed in my inbox this week — resulting from a study surveying over 1,500 musicians on their AI use — I was stunned.
AI remains a hugely divisive topic among music producers, with a significant chunk still very much sceptical about its place in music. Damon Albarn, for example, who is currently writing the score for a movie about OpenAI, recently said it “isn’t possible for AI to make soulful music”.
There are also concerns about AI-generated music flooding streaming platforms, with a November 2025 study from Tracklib finding that over 80% of producers are against AI-generated songs.
So is it really the case that 78% of professional musicians are now using AI in their workflows? No, not quite.

READ MORE: I tested 9 of the best stem separation tools — here’s how they compare

The new study – conducted by AI-powered music platform Moises in partnership with music research firm Water & Music – concludes that it challenges “prevailing narratives about AI” and those who say AI might not be the panacea it’s sold to be.
But the small print is that 80% of the 1,525 surveyed musicians are customers of Moises. The company is built around an AI-focused music-making platform, with context-aware AI, stem separation powered by AI, stem generation with AI and AI voice conversion, with many other similar features.
This looks like a case of selection bias – those using an AI music platform are already comfortable using AI tools, and thus already have positive attitudes towards them. So, the study was framed – not necessarily intentionally – for results which skewed in favour of AI use among music producers.
A more accurate conclusion to draw would perhaps be: “Among musicians already in the AI ecosystem, 78% are regularly using AI in their work.”
So why does this matter? Moises is an AI company, and will naturally benefit from narratives showing high AI adoption rates. Human behaviour is infectious; if we see more of our peers adopting a particular behaviour, we’re more likely to follow suit. Effectively, it might be placing undue pressure on producers to adopt AI to keep up.
It may also lead investors in AI technology to be overconfident, and those in charge of AI regulation to downplay concerns about copyright and training data, assuming resistance is limited rather than widespread, and consequently making ill-informed policy decisions.
Moises logo. Credit: Moises
The Moises study isn’t the only study showing widespread AI adoption rates among producers, but conducted by a company for which AI adoption is a key motivator.
In November 2025, AI-driven audio mastering platform LANDR released its own report, which found that “87% of artists have incorporated AI into at least one part of their process”. It’s just that the 1,241 respondents were sampled from “LANDR’s global community” – existing users of an AI platform. I’m not entirely sure how the remaining 13% is accounted for – are they people who have signed up to LANDR but never used its products?
In a world placing ever-increasing pressure on news outlets to be the first to cover breaking stories, it’s all too tempting to take the headline of a press release and run an article without thoroughly questioning the data or information supplied.
A study made headlines in February when it found that of 81 popular headphone models from Bose, Samsung, Sennheiser and more, all contained hazardous, in some cases “feminising” and even cancer-causing chemicals. The study – titled The Sound of Contamination – was published by Czech non-profit environmental organisation Arnika, as part of the EU-funded ToxFree LIFE For All initiative.
The findings have led some European retailers to begin pulling headphone models from the market, but spokespeople for some of the headphone brands have questioned the legitimacy of the study.
“The study used its own testing criteria and flagged the product based on thresholds for BPA-related substances that are stricter than those typically applied to plastics used in electronic products,” said Anna Forsgren, product compliance and sustainability manager at Marshall Group.
Elsewhere, Sennheiser spokesperson Eric Palonen revealed the company contacted authors of the report “hoping to get the exact data for the Sennheiser products tested in order to verify our data and decide on next steps,” but added authors did not provide the requested data.
It’s a noble pursuit to ensure people aren’t being poisoned by their headphones. But when a study makes retailers actively pull products, and sows widespread worry among headphone users, it’s important that the conclusions are being drawn from accurate and non-biased data, and that the conditions through which the study is set up are statistically sound.
AI adoption is no doubt increasing among music products, with swathes of useful new tools constantly hitting the market, offering improvements to production workflows. But it’s worth being vigilant and not taking all studies as gospel – there’s no pressure to adopt new technologies into your workflow if you don’t actually need to
You can try out Moises’ AI-powered stem separation and voice conversion features by downloading it for free at the official Moises website.
The post Study shows “78% of musicians are now using AI” – but is everything really as it seems? appeared first on MusicTech.