Stimulate, Don't Replace: How AI Design Shapes Human Creative Effort. The CineCoach Case and Recommendations for European Policy

Films from the 27 Member States of the European Union, counted with their predecessor states, have won 52 of the 79 Academy Awards for Best International Feature Film and are credited on 54 of the 102 winners of the Palme d'Or. The festivals of Venice, Cannes and Berlin are European. Some 8.9 million Europeans work in culture in 24 official languages. Generative AI now touches every one of these fields. The European Commission is preparing an AI strategy for the cultural and creative sectors on the principle that AI should complement and not replace human creativity. We ask what the evidence says about how that principle can be built into a tool and what it implies for public policy. We review seven bodies of research (neural and cognitive models of creativity, the generation effect, motivation and evaluation, deliberate practice and feedback, tutoring and Socratic instruction, guided discovery and human–AI co-creation) and report only findings from sources we opened with their statistics as stated. Human–AI co-creation raises individual output and lowers collective diversity (meta-analytic g = 0.27 and g = −0.86 with a second meta-analysis putting homogenisation at d = 0.33). The interface decides which effect dominates – an AI that questions the person about their own idea preserves diversity and ownership while an AI that rewrites the idea loses both (d = 0.76 and 0.57). In learning an answer-giving AI raises assisted performance and lowers unassisted performance (−17%) while a hint-giving AI does no harm. High-information feedback has twice the effect of corrective feedback (d = 0.99 against 0.46). We also report the findings that cut against a simple Socratic thesis – a randomised trial in which a strategy-giving AI coach left users worse off than no AI at all, another with no difference between Socratic and non-Socratic tutors and the absence of any reliable neural signature for divergent thinking. Used wrongly AI will flatten the plurality that makes European creativity unique. Used rightly it will augment and accelerate European creativity. We describe CineCoach (a coaching system encoded from one cinematographer's method and running in fifteen languages, eleven of them official EU languages, each with its own natively analysed film curation) as one working example of the augmenting design and we state its limits. We close with an agenda for the strategy – fund the design rather than the technology, measure output diversity as an outcome and put artist-facing AI under the control of the schools and studios that use it.