Academic researchers from Deakin University and Swinburne University of Technology, working with data from TBWA\Australia’s DISRUPT AI Film Festival, have found that while generative AI is reducing creative production costs, human judgement remains the critical differentiator in the quality of creative output.
The white paper, titled Creative AI as National Infrastructure, draws on 371 festival submissions, jury deliberations, and participant data from what the release describes as Australia’s first dedicated AI film festival. It finds that creators are producing narrative work at a fraction of traditional costs — several participants completed films for under A$1,000 — but that lower production costs did not correlate with stronger creative outcomes.
The most compelling films consistently demonstrated three human capacities: clear intention, discerning taste, and critical judgement.
“As AI makes production cheaper, the scarce skill becomes knowing what is worth making,” said Lucio Ribeiro, Chief AI & Innovation Officer, TBWA\Australia.
“That transforms judgement, taste and intention from soft creative traits into workforce capabilities. Australia can either train tool operators or build creative AI practitioners who understand craft, support better decision-making and build communities of practice.”
The creative advantage is shifting
As AI tools become widely available, the research argues that technical access alone no longer separates strong work from mediocre. What matters, according to the paper, is craft — knowing when to use AI, when to stop, what to reject, and how to maintain creative purpose through rapid iterations.
For brands, agencies, and creators, this reshapes where value concentrates: production accelerates, visual exploration becomes cheaper, and more people can execute work that previously required large teams and budgets.
A risk the paper identifies is technically impressive but culturally hollow output flooding the market.
“The strongest DISRUPT creative films had something crucial: a point of view, restraint, and intentional decision-making,” Ribeiro noted. “Better tools alone don’t create better work.”
A local community is forming, but unevenly
Australia has a creative AI community forming, though it remains geographically concentrated. DISRUPT was designed as both a research platform and training ground, supporting creators learning through experiments, informal networks, meet-ups, online communities, and self-directed practice.
The research found the average Australian participant was 44.7 years old, with the largest cohort aged 40 to 49. Only six of 73 Australian submissions came from students, which the paper identifies as a gap between tertiary screen education and current generative AI practice. Participation was concentrated in Melbourne and Sydney. Among first-time filmmakers, however, female participation reached 46%, compared with 16% across all entrants.
“The diversity showed that as barriers fall, participation changes,” Ribeiro said. “When creative tools democratise but industry structures don’t, the next generation of talent may emerge through entirely different pathways.”
The tool question is also a local value question
Most major tools used by DISRUPT participants were owned by U.S. or Chinese companies, reflecting what the paper describes as the current generative AI ecosystem reality: Australian creators build skills and IP on global platforms. The festival’s partnerships with Google and Leonardo.ai are presented as a model for using international infrastructure to support Australian community-building and local IP development.
The research identifies the governance of First Nations stories and cultural knowledge as a particular area requiring attention, stating that governance must remain on First Nations terms.
Implications and recommendations
The white paper identifies three imperatives for brands and creative leaders: investing in craft and judgement rather than tool access alone; supporting Australian creative AI communities through decentralized practice networks; and protecting cultural specificity against the risk of generic global outputs.
Six formal recommendations are proposed: a Creative AI Industry Development Policy; generative AI filmmaking in screen education by 2027; clearer authorship and attribution standards; sovereign creative AI tooling research; an Australian Creative AI Futures Index; and support for DISRUPT as a national knowledge-exchange platform.
“When we launched DISRUPT in 2025, we wanted to test one of the biggest conventions in creative AI: that better tools automatically create good work faster,” Ribeiro said.
“Across 371 films, evidence pointed elsewhere. AI helps, but the strongest work still depends on intention, taste and critical judgement.
“This research gives Australian brands and creators a more practical framework for navigating AI with sharper judgement, stronger cultural awareness, and clearer vision of where creative value is moving.”
Download the full white paper here.
