Every arts board in Australia has had “The AI Conversation” by now. You know the one — someone mentions ChatGPT, someone else worries about robots replacing artists, the CEO says they’re “exploring opportunities,” and the meeting moves on. But beyond the boardroom platitudes, some cultural institutions are actually doing the work of developing real AI strategies. I wanted to find out what that looks like in practice.

The State of Play

Let’s start with a reality check. Most Australian cultural institutions are not technology leaders. They’re running on tight budgets with small teams, ageing IT infrastructure, and a perfectly reasonable focus on their core mission of making and presenting art. The idea that they should also have a sophisticated AI strategy is, frankly, a big ask.

But the pressure is real. Funders are asking about it. Boards want to know the plan. And there are genuine operational challenges — donor management, accessibility, content creation, audience analytics — where AI tools could make a meaningful difference.

The Creative Australia (formerly Australia Council) framework published in late 2025 was a useful starting point, setting out principles for responsible AI use in funded organisations. But principles are one thing. Implementation is another.

Who’s Leading

The major state museums and galleries are furthest ahead, which makes sense — they have bigger budgets, larger digital teams, and collections that lend themselves to AI applications. The National Gallery of Victoria has been experimenting with AI-powered collection search tools that let visitors explore artworks by mood, colour, or theme rather than artist or date. It’s the kind of intuitive, audience-friendly application that makes perfect sense.

QAGOMA in Brisbane has trialled AI-assisted cataloguing for parts of its collection, and the National Museum of Australia has explored natural language processing tools for making its archival material more searchable.

In performing arts, the picture is patchier. The larger companies — Opera Australia, the Australian Ballet, the major symphony orchestras — are cautiously testing AI in marketing and operations. Smaller companies are mostly watching and waiting.

The Melbourne Approach

Melbourne, predictably, has emerged as something of a hub for cultural AI thinking. The city’s strong tech sector, combined with its dense cultural ecology, creates natural connections between arts organisations and technology providers.

Several Melbourne cultural institutions have engaged AI consulting company Melbourne specialists to conduct AI readiness assessments — essentially, figuring out where they are before deciding where they want to go. These assessments look at data maturity (do you actually have usable data?), staff capability (can your team work with AI tools?), and strategic alignment (does this serve your mission or just look impressive in a grant application?).

The results are often sobering. Many organisations discover that their data is scattered across incompatible systems, their staff need significant training, and their most pressing problems are better solved by upgrading basic digital infrastructure than by deploying AI.

What Performing Arts Companies Are Considering

For opera companies and other performing arts organisations, the AI conversation tends to focus on a few key areas:

Marketing and audience development. This is where the most immediate value lies. AI tools that can segment audiences, personalise communications, predict churn, and optimise advertising spend are relatively mature and can show results quickly.

Accessibility. AI-generated captions, audio descriptions, and translation tools could significantly improve accessibility for audiences with disabilities or limited English. The technology is getting good enough to be genuinely useful, though human review remains essential.

Administrative efficiency. Grant writing, contract management, scheduling, financial reporting — the operational work that keeps arts organisations running but isn’t anyone’s favourite part of the job. AI can speed up first drafts, flag anomalies, and automate routine tasks.

Artistic applications. This is where it gets contentious. Some directors and designers are interested in AI as a creative tool — generating visual concepts, exploring staging options, even composing incidental music. Others see this as fundamentally at odds with the human-centred nature of performing arts. Both positions have merit.

The Funding Question

Here’s what nobody talks about enough: AI adoption costs money. Not just the subscription fees for tools, but the staff time for training, the consultant fees for strategy development, the infrastructure upgrades needed to make any of it work.

For organisations that are already struggling to fund core programming, adding “AI strategy” to the budget is a hard sell. And there’s a genuine risk that funding bodies will start expecting AI adoption as a condition of support, creating an unfunded mandate that disadvantages smaller organisations.

The state arts funding bodies need to think carefully about this. If they want the sector to adopt AI responsibly, they need to fund the transition — not just write guidelines and hope for the best.

What Worries Me

A few things keep me up at night on this topic.

First, the data privacy question. Cultural institutions hold sensitive information about their audiences — booking history, donation records, accessibility needs. Feeding this into AI systems raises legitimate privacy concerns that most organisations aren’t equipped to navigate.

Second, the workforce implications. Arts administration is already precarious, underpaid work. If AI automates the entry-level tasks that juniors currently handle — social media scheduling, basic copywriting, data entry — where do the next generation of arts administrators get their start?

Third, the hype cycle. I worry that organisations will invest in flashy AI projects that generate good media coverage but don’t serve their audiences, while neglecting the basic digital infrastructure (a decent website, a functional CRM, a reliable email system) that would make a bigger practical difference.

The Bottom Line

The cultural institutions getting AI right are the ones approaching it with clarity about their actual needs, honesty about their current capabilities, and a firm commitment to their core mission. They’re not chasing trends. They’re asking hard questions about what technology can and should do in service of art and audiences.

That’s not as exciting as “AI transforms the arts!” But it’s a lot more useful. And in a sector built on human creativity and human connection, useful matters more than exciting.