The Innovation Lab: OnstageAI x Malmö Opera
- Paulina Pomykała
- 2 days ago
- 3 min read
Interview with Jesper Larsson, Head of the Production, Planning, and Audience Section at Malmö Opera, Sweden
Heading into 2026, we are beginning a collaboration with Malmö Opera (Sweden), an institution known for its willingness to experiment where art and technology meet. We are working closely with the Malmö team to refine our autonomous systems and develop features tailored to the specific demands of opera, such as subtitle automation. To provide a perspective from the house's side, we sat down with Jesper Larsson, Malmö’s Head of Production, Planning, Marketing and Sales, to discuss their history of innovation and where they see AI heading next.
Malmö Opera has been at the forefront of this space with the Kalaudioscope project. What were your key takeaways from that experience regarding the use of AI in a live performance environment?
The idea behind Kalaudioscope is to explore possibilities for individualised audience experiences of the live performance in a digital space. Just as the experience in the physical space is different to everyone in relation to how it sounds and looks dependent on where you sit in the auditorium or choose to look at any given time, we want to offer the audience the same, or even broader, tailored content. In the existing streams we offer every digital audience the same package for everyone which we are trying to find ways to change. We have been working on this now for three years in an innovation portfolio consisting of seven areas: video, audio, UX, datastream/synch/cloud, legal issues, business models and new artistic formats. In all of the areas there are opportunities to use AI in different forms but a lot also need to be done on a policy level at the same time.
We are currently exploring the automation of surtitles in collaboration with you. From an opera house's perspective, what are the biggest opportunities and potential challenges?
In an opera house everything is related to the music, the tempi and flow of it. All cues for scene changes, entrances, light changes, surtitling etc. are all related to where we are in the music and a lot of manual systems are operated to follow the flow of the music to be able to do the right thing at exactly the right time for precisely this performance. The opportunities with a system that is able to accurately determine, in real time, where we are in the score are massive. We are excited to collaborate with OnstageAI to explore what and how this could be done. The challenges with this is of course to prove that the system is reliable, precise and without latency, and doesn’t mean increased risk of failure or the need to be monitored by yet another manual handling.
You’ve mentioned before that technology should help 'democratize' the opera. Looking ahead five years, how do you think AI-driven features will change the audience experiences and internal processes?
Wow! I think that the audience will have individualised help from AI for both the on-site, physical experience and for new digital versions. AI will help making sure that every audience member are provided tailored information and preparation before, during and after the visit which will both enhance the experience and bring more people to the performances. This could be in every field from marketing and sales to introductions, connection to the artists, digital individualised performance programs, AR features in the foyer or even during the performance and after talks. BUT also thinking of the digital twin of the opera house where a parallel experience could be available to everyone, regardless of physical ability, geography, language, economy etc. If this experience offers more than today’s televised versions and at the same time demands less resources to produce it will result in more people getting access to the arts and more potential incomes for producers and artists. Internally in any arts organisation we can use AI to help us automate a huge part of the administrative work as well as combining systems that are necessary for the performances. If we reduce the number of tasks that are manually handled putting the same information into different systems we can spend the time on making more art.