There’s been a lot of breathless coverage about artificial intelligence transforming the arts. Most of it, frankly, has been written by people who’ve never sat through a four-hour Wagner cycle or tried to coordinate 200 performers for an opening night. So let me bring some perspective from someone who has.

Australian arts organisations — from major opera companies to mid-tier theatre groups — are experimenting with AI. But the reality is far more measured, and far more interesting, than the headlines suggest.

What Opera Australia Is Actually Doing

Opera Australia has been quietly trialling AI tools in a few behind-the-scenes areas. Nothing dramatic. Think automated captioning for rehearsal recordings, AI-assisted translation drafts for surtitles, and some early experiments with predictive analytics for ticket sales.

The surtitle work is genuinely useful. Anyone who’s worked on a new production knows that translating a libretto for surtitles is painstaking — you need to convey meaning in short phrases that audiences can read at a glance while watching the stage. Having AI generate a first draft that a human translator then refines? That saves real time and money.

The ticketing analytics are more interesting still. By analysing historical booking patterns, weather data, and even social media sentiment, companies are getting better at predicting which performances will sell out and which need a marketing push. It’s not revolutionary — retail has done this for years — but it’s new for the arts.

The State Companies Are Watching

The state opera companies are at different stages. Victorian Opera has been experimenting with AI-generated marketing copy, though artistic director Stuart Maunder has been clear that creative decisions remain firmly human. State Opera South Australia has looked at AI tools for donor relationship management, which sounds dry until you realise that donor revenue can make or break a season.

West Australian Opera, with its smaller team, sees AI as a way to do more with less — automating grant application formatting, generating first drafts of program notes, that sort of thing.

Where It Gets Uncomfortable

Here’s where I put on my opinionated hat. There’s a real tension between efficiency and artistry, and not every arts organisation is navigating it well.

I’ve heard of at least one major festival considering AI-generated visual art for marketing materials instead of commissioning local artists. That’s a problem. The arts sector exists to support human creativity. Using AI to replace the very artists you’re supposed to champion is, at best, tone-deaf.

The musicians’ union has also raised concerns about AI-generated orchestral demos being used in pitch meetings. When you’re trying to convince a board to fund a new commission, playing an AI-rendered version of the score instead of hiring actual musicians sends a troubling message.

The Brisbane Connection

Interestingly, some of the more innovative work is happening in Queensland. A few arts organisations there have engaged AI consultants in Brisbane to audit their operations and identify where AI might genuinely help without undermining artistic values. The key word is “audit” — looking before leaping, rather than just throwing ChatGPT at everything.

Brisbane’s QPAC has been particularly thoughtful about this, establishing clear guidelines about where AI tools can and can’t be used in their operations.

What I Think Is Worth Watching

The most promising applications aren’t the flashy ones. They’re the boring operational improvements that free up humans to do what humans do best.

AI scheduling tools that can juggle rehearsal room availability, singer conflicts, and union requirements? Genuinely useful. AI that analyses acoustic data to help sound engineers optimise a venue? Interesting. AI that writes the opera itself? Not interested, thanks.

The Australia Council for the Arts released guidance last year on AI use in funded organisations, and it struck a reasonable balance: embrace tools that support artistic work, but maintain transparency about what’s AI-generated and what isn’t.

The Bottom Line

Australian arts organisations are being cautious, and I think that’s wise. The worst outcome would be rushing to adopt AI tools that compromise the thing audiences actually come for — the irreplaceable experience of live, human performance.

We’re in the early days. The organisations getting it right are the ones asking “does this help us make better art?” rather than “does this make us look innovative?”

I’ll be watching this space closely. And yes, I wrote every word of this myself.