news

Adobe's creator survey leads with 78% growth, and reports 68% settling for AI output that misses

Both figures come from the same self-report of more than 2,000 US creators polled by The Harris Poll in October 2026, with no measurement of the output itself.

By Daniel Ochoa · 11 November 2026

Adobe's creator research, published on November 9 ahead of MAX 2026, carries two numbers that sit awkwardly together. Of creators who currently use creative AI, 78 percent say it has accelerated the growth of their business or following. In the same survey, 68 percent say they at least sometimes settle for a result that does not quite match what they had in mind. Both are self-reported, and neither measures a single generated image or video.

The figures Adobe published, in Adobe's own order

The headline Adobe chose is the 78 percent growth figure. Further down the same post sit the rest: 60 percent say creative AI has made it at least somewhat easier to produce work that stands out from other creators, 60 percent believe creators with advanced AI skills are getting opportunities they are not, 51 percent expect that gap to widen, and 89 percent of current creative AI users are confident they can learn the AI skills they need. The 68 percent who accept off-target output appear under a heading about settling for close enough. Trade coverage of the MAX announcements added a further figure from the same research: 77 percent said the final creative decision should always remain with a human creator.

The methodology line is short and specific. Adobe worked with The Harris Poll and surveyed more than 2,000 US creators in October 2026. That is a self-assessment panel, not an evaluation of the files those creators produced, so the 68 percent is a statement about perceived shortfall rather than a measured error rate.

Why a shortfall figure is not an output score

Reported dissatisfaction and measured defect rates answer different questions. A creator who settles for close enough may be describing a composition that drifted from a brief, a hand with the wrong number of fingers, text on a sign that rendered as noise, or simply a deadline. Nothing in the survey separates those, and nothing in it ties the shortfall to a named model. Per-category scoring does separate them, which is why we score the same prompt set across models and publish the categories individually rather than reporting one satisfaction percentage.

The sample is not the same one Adobe used in June

Adobe released a different piece of creator research earlier in 2026, its 2026 Creators' Toolkit Report, published on June 16 and based on a survey of more than 16,000 creators across Australia, France, Germany, India, Japan, South Korea, the UK and the US. The November research covers more than 2,000 creators in the US only. The two therefore should not be read as consecutive points on a trend line: the geography, the sample size and the fielding date all changed between them.

That earlier round attracted criticism for how narrowly it defined a creator, with coverage at the time arguing that the panel skewed heavily toward social media creators rather than the wider professional creative industry. Adobe has not published a respondent breakdown for the October 2026 round alongside the headline percentages, so the same question applies to the 78 and 68 figures now in circulation during MAX week.

For a company that sells both the generators and the editors used to fix their output, the most load-bearing number in the release is the one about settling. It is also the one with the least evidence behind it.

Sources: Adobe Blog, TheWrap, ITBrief, 9to5Mac, AppleInsider.

Daniel Ochoa: Covers model launches, shutdowns and pricing changes as they happen. Reads deprecation notices for a living so you do not have to.