Nornic Register · Living document
“AI is causing a collapse in creative effectiveness.”
The study most often cited for this predates generative AI entirely. There may be a real effect; this is not evidence of it.
The source
- Study
- The Crisis in Creative Effectiveness
- Publisher
- IPA (Institute of Practitioners in Advertising)
- Publication type
- Industry analysis
- Sample
- IPA Databank cases
- Fieldwork
- Published October 2019, analysing 1996–2018 data
- Funding & conflicts
- IPA — an advertising industry body
- Verified on
- 2026-07-26
The arithmetic
The analysis covers 1996 to 2018 and was published in October 2019. Generative AI tools were not in mainstream creative production in that window.
Whatever the study found, it cannot be evidence about the effect of a technology that arrived afterwards.
A separate, genuinely relevant finding does exist: a meta-analysis of 19 studies and 61 effect sizes reports statistically significant homogenisation of creative output under generative AI, d = 0.334. That is about sameness, not effectiveness.
The two are often conflated. Convergence and ineffectiveness are different claims requiring different evidence.
What would change this verdict
A study measuring campaign effectiveness against generative-AI usage over a post-2022 window, with the confound of budget and category controlled.
Sources