If you’ve been to an opera in the last thirty years, you’ve probably relied on supertitles — those projected English text lines above the stage that tell you what’s being sung. They changed opera for audiences who don’t speak Italian, German, French, or Czech (which is most of us, let’s be honest). But the way supertitles are created and delivered hasn’t changed much since the 1980s. That’s starting to shift.

Several opera companies internationally are now experimenting with AI-powered supertitle systems, and the implications for how we experience opera are significant.

How It Currently Works

Right now, supertitles are prepared weeks before a production opens. A translator creates the English text, then a supertitle operator sits in a booth during every performance and manually triggers each line in sync with the singers. It’s a skilled job — you need to know the score intimately, anticipate the conductor’s tempo choices, and react in real time when a singer holds a note longer than expected or skips a repeat.

I’ve watched supertitle operators work, and it’s genuinely impressive. But it’s also expensive (you need an operator at every performance), inflexible (the translations are fixed), and limited to one language at a time.

What AI Changes

The new systems use a combination of automatic speech recognition, natural language processing, and pre-loaded score data to generate and display supertitles with minimal human intervention. Some systems go further: they can offer translations in multiple languages simultaneously, delivered to individual seat-back screens or audience members’ phones.

The Royal Opera House in London has been trialling a system that offers supertitles in twelve languages. The Bavarian State Opera in Munich has tested AI-assisted translation for lesser-performed works where professional translations don’t already exist. And several American companies have been experimenting with real-time captioning for accessibility — not just translation but genuine closed-captioning for deaf and hard-of-hearing audience members.

The Australian Angle

In Australia, arts organisations are generally a few years behind the major European and American houses when it comes to technology adoption. But there’s growing interest. I spoke with a technology consultant who works with several performing arts organisations, and they noted that companies are increasingly engaging AI consultants Melbourne to assess where AI tools can improve audience experience without blowing already-tight budgets.

The obvious win for Australian companies is multilingual supertitles. Australia’s opera audiences are increasingly diverse. In Sydney and Melbourne especially, significant portions of the audience speak Mandarin, Cantonese, Korean, Vietnamese, or Hindi as a first language. Offering supertitles in those languages — even via a phone app synced to the performance — could be genuinely transformative for audience development.

Opera Australia already offers audio-described performances for vision-impaired audiences. Adding multilingual supertitles would be the next logical accessibility step.

The Trade-Offs

I’d be a bad critic if I didn’t flag the concerns.

Quality of translation. Opera translation is an art form in itself. A good supertitle translation doesn’t just convey meaning — it captures rhythm, poetry, humour, and subtext in a way that complements the music. AI translations are getting better rapidly, but they still struggle with the nuance that makes opera text special. A line like Violetta’s “Amami, Alfredo” is simple to translate literally (“Love me, Alfredo”) but a good supertitle translator knows to time that line with the musical climax in a way that amplifies the emotional impact. AI doesn’t understand musical dramaturgy. Not yet.

Screen glow. If audience members are reading supertitles on their phones, that’s dozens of glowing screens in a darkened theatre. Anyone who’s been to a movie lately knows how distracting this is. Seat-back screens are better but expensive to install.

Job displacement. Supertitle operators are skilled professionals. If AI replaces them, that’s another specialist role lost in an industry that already employs too few people.

Where I Land

I’m cautiously optimistic. The accessibility benefits are real and important. If AI-powered supertitles can help a Mandarin-speaking family in Hurstville experience La Traviata in their own language, that’s worth pursuing. If real-time captioning can make opera genuinely accessible to deaf audiences, that matters enormously.

But I’d want human oversight of the translations, I’d want phone screens banned in favour of seat-back or seatback-app solutions with dimmed displays, and I’d want the companies to commit to retraining rather than simply firing their existing supertitle teams.

The technology is coming regardless. The question is whether opera companies will implement it thoughtfully or just grab the cheapest option and hope for the best.

Given the industry’s track record with technology adoption… I have opinions. But I’ll save those for another column.

— Margot