AI is slashing the cost of creative production, but according to new research, human judgement is still the only thing that actually makes it good.
Research from Deakin University and Swinburne University of Technology, drawing on submissions to TBWA\Australia‘s DISRUPT AI Film Festival, revealed that while generative AI is fundamentally changing creative production economics, human judgement remains the critical differentiator.
The white paper, ‘Creative AI as National Infrastructure’, analysed submissions, jury deliberations and participant data from Australia’s first dedicated AI Film Festival. Findings showed the most compelling films demonstrated three human capacities: clear intention, discerning taste, and critical judgement.
TBWA\Australia chief AI and innovation officer Lucio Ribeiro said: “As AI makes production cheaper, the scarce skill becomes knowing what is worth making,”
“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 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.”
Australia has a creative AI community forming, though it remains geographically concentrated and decentralised. DISRUPT was designed as both research platform and training ground to support creators learning through experiments, informal networks, meet-ups, online communities and self-directed practice.
Research revealed insights about participation patterns. The average Australian participant was 44.7 years old, with the largest cohort aged 40-49. Only 6 of 73 Australian submissions came from students, indicating a gap between tertiary screen education and current GenAI practice. Participation concentrated in Melbourne and Sydney, where creative networks and technical confidence are strongest.
However, data showed promising signs: among first-time filmmakers, female participation reached 46 per cent, compared with 16 per cent 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.”
Most major tools used by DISRUPT participants were owned by US or Chinese companies, reflecting the current GenAI ecosystem reality that Australian creators build skills and IP on global platforms.
The opportunity lies in ensuring global technology strengthens local capability. DISRUPT’s partnerships with Google and Leonardo.ai demonstrate this model, international infrastructure supporting Australian-born tooling, community-building and local IP development together.
The research identified three key imperatives:
Invest in craft and judgement, not tool access alone. As AI reduces execution costs, better decision-making and technical fluency become valuable capabilities.
Support Australian creative AI communities. The field develops through decentralised practice, not formal institutions alone. Brands and agencies can facilitate learning networks where practitioners share and promote local knowledge.
Protect cultural specificity. Generic global outputs risk creative sameness. Work grounded in local context and intent has greater impact.
Ribeiro said: “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.”
