Sydney Opera House sets expectations for AI – Emerging Tech

The Sydney Opera House has laid out its thinking around AI use, with a desire to improve its online experience while still taking a cautious approach to the emerging technology.

Chief financial officer Jon Blackburn told a Gartner CFO and Finance executive conference that as a government organisation it must be cautious, but opportunities exist.

The multi-venue performing arts centre has already undergone a five-year digital transformation journey, which saw it shift from legacy platforms to the cloud.

It runs an Oracle-based budgeting and forecasting system, and is talking to the vendor for some proofs-of-concept, including around predictive analytics.

Blackburn said that when it comes to AI, the Sydney Opera House is not in a position to be an early adopter, and would take a risk-based approach to adoption.

A key risk identified with AI use is the potential for corporate data to find its way into the public domain.

“We didn’t want Opera House data leaving [the organisation] – we prevented people from using ChatGPT and doing stuff externally [for that reason],” Blackburn said.

However, Blackburn said he is open to the idea of adopting AI that comes “embedded” into its existing systems and license conditions, as long as it is “not … an additive cost to the business”.

For specific licensing of AI software, Blackburn outlined a preference to apply it only in specific areas on a needs basis.

“With AI and some of the AI agents, you can bring them to the areas that you need some help in that are repetitive, that can benefit from that, and you’re paying [for] a license just for those people to use it, rather than having to provide everyone in the organisation with access to AI,” he said.

Enabling widespread AI access, on the flipside, could increase the Opera House’s risk.

He did not think M365 Copilot would work for the Opera House, in part over fears that the agent might “get into places that it shouldn’t” when looking for data or responses.

“Corralling where the AI can go and what it can do, and putting it into use cases, is where we are now doing it,” Blackburn said.