Do creatives actually want AI?

Three surveys this year tell the same story from different angles: almost everyone in creative work is using AI, and very few are convinced by it. Luma's survey of US creative professionals found 81% have released AI-made or AI-assisted content, but only a minority say it improves quality as well as speed. Creative Boom's survey found 86% use AI while just 10% think it is good for the industry. Research on designers found weekly use became near-universal within a year. The tools won on availability. The argument is now about where they belong. What the numbers describe is not a verdict on the tools but a gap between using something and being persuaded by it.

Luma's report on creative professionals and AI is based on a survey of US creatives across agencies, brands, production companies and freelance creators, and its shape is the one people describe in conversation: heavy use, thin confidence. 81% have released AI-generated or AI-assisted content, and most rate their access to the tools as good. But only a minority say the tools improve creative quality as well as efficiency, many are concerned about over-reliance, and more than half work somewhere with no formal process for evaluating or approving AI tools. The same organisations tried more tools than they kept, and dropped a good number along the way. It is a company-commissioned survey of one market, so we read it as an industry snapshot rather than a verdict — but the gap between using a tool and trusting it is exactly what people tell us.

Creative Boom's survey of the creative industry pushes the point further, and it is harder to dismiss because its question is not about use but about effect. 86% use AI tools in their work, yet only 10% think its overall effect on the industry is positive; most call the impact mixed or outright negative. The same survey found widespread burnout and a sense that people are less financially secure than a year earlier. The two facts sit together without contradiction: people are adopting the tools because clients and timelines expect them, not because they have been persuaded. We would read the mood as a warning about the terms of adoption, not about the technology itself.

Research on designers tracks the same year in practice rather than in attitude. Weekly use of AI in design work jumped within a year, daily use followed, and the average designer's list of off-the-shelf tools doubled. The same research found that half had pushed AI-generated code into production, and it records concerns about craft fading and about the isolation of designing alongside a model rather than teammates. That is the trade the numbers describe: more throughput, more tools, and a profession quietly asking what the throughput is for. What the research measures is habit rather than belief: what people reach for on an ordinary working day, and how quickly the reaching becomes routine.

None of these three surveys covers Singapore, and we will not present them as if they did. Luma's is US-only, fielded for a US brand conference; Creative Boom's is weighted to the UK and the United States; the designer research is mainly the United States. We have no comparable local study of how Singapore's designers, editors and studios feel about their own AI use, so we are not going to borrow someone else's percentages and call them ours. If your team is deciding on this, the honest position is that the shape of the story travels — near-universal use, low confidence — while the numbers do not. That is a gap we would rather name than paper over with a borrowed figure.

We do not read the 10% who think AI is good for the industry as stubbornness. It is what a market looks like when the tool arrives before the standards do: everyone adopts, nobody has agreed on disclosure, pricing or authorship, and the people doing the work can feel the difference between making something and finishing something. We would trust least the survey that measures enthusiasm; the numbers on tool counts and weekly use describe what has already happened.

What we would take from it is narrower and more practical. Decide where AI is allowed to touch the work before a client asks, keep the style and character decisions with a person, and be honest with clients about what was made how. The surveys say the tools are in the room now. They do not say anyone has worked out the manners.