Review: AI Talk live podcast shows the breadth of the Canton of Zurich's AI sandbox

It was precisely this composition that made the format so appealing: Perspectives from clinical practice, HealthTech entrepreneurship, language AI and cantonal innovation promotion came together on stage. The evening thus showed in an exemplary manner that the AI sandbox is not an initiative for a single sector, but a platform on which AI applications are tested under real conditions, classified in regulatory terms and further developed in the interplay of technology, administration and practice.

Raphael von Thiessen and the sandbox perspective

Raphael von Thiessen, who heads the Innovation Sandbox for AI at the Canton of Zurich, played a key role. The Innovation Sandbox describes itself as a test environment in which players can implement AI projects within a clearly defined framework if technological, regulatory or data-related uncertainties would otherwise slow down innovation. The Sandbox aims to promote responsible AI innovation, build expertise in AI regulation and strengthen the interdisciplinary ecosystem in the Zurich area.

The fact that Raphael von Thiessen was present on the panel lent the event an important level of classification. It stood not only for the perspective of the Division of Business and Economic Development, but also for the question of how individual pilot projects can generate systematic findings for administration, business and research.

Gábor Márk Somfai and the Digital Eye Clinic

Gábor Márk Somfai, a representative of clinical practice who works as a senior physician at the Zollikerberg Eye Center and the Eye Clinic of Zurich City Hospital, was on stage. The Digital Eye Clinic project he represents stands for AI-supported diagnostics and the question of how image-based procedures can be reliably integrated into medical processes.

This project in particular illustrates how AI can be used in highly sensitive medical contexts without the requirements for validation, responsibility and trust becoming less important. During the live podcast, he made it clear that medical innovation is not only measured by technological performance, but also by how well it fits into everyday clinical practice.

Michelle Plüss and medical documentation

With MPAssist, Michelle Plüss contributed the perspective of a HealthTech start-up that deals with AI-supported medical documentation. The official event profile describes its contribution as an approach that relieves doctors, automates medical reports and at the same time increases quality and efficiency.

The focus was therefore on a field of application that is particularly tangible for many institutions: administrative relief in the healthcare sector. At the same time, this field in particular shows how closely productivity, data protection, professional secrecy and regulatory issues are intertwined as soon as AI works with sensitive health data.

Paula Reichenberg and machine translation

Paula Reichenberg represented the field of machine translation with NEUR.ON and spoke about the use of modern translation models in specialized contexts. The event profile emphasizes that quality, context and ethics are crucial if language barriers are to be overcome with AI.

It is important for the classification of the evening that this project is not assigned to the medical field, but to the legal or specialized institutional environment. It was precisely this difference that made it clear how broadly the sandbox is designed: it includes not only HealthTech, but also applications in which language, precision and institutional responsibility are at the forefront.

Three projects, one common logic

The three projects presented differed significantly in their technical logic: clinical diagnostics, medical documentation and machine translation in a legally sensitive context. Nevertheless, they were united by a common basic question: How can AI be developed and used under real conditions in such a way that it creates concrete benefits and at the same time meets the requirements of security, transparency and responsibility?

This is where the strength of the Innovation Sandbox becomes apparent. It is not just a showcase initiative for futuristic use cases, but a framework in which companies, research and the public sector work together to transform AI from an idea into viable practice.

More than just a technology event

The AI Talk Live podcast was therefore much more than a discussion about tools or trends. He showed that a productive approach to AI begins where different roles come together: a moderator who structures the debate in an understandable way, a sandbox manager who sets out the institutional framework, and guests who report on their specific projects.

With Stephan Lendi as host, Raphael von Thiessen as the voice of the Innovation Sandbox and Gábor Márk Somfai, Michelle Plüss and Paula Reichenberg as project representatives, a format was created that made the breadth of the Zurich AI ecosystem visible. It was precisely this combination that sent out the message of the evening: AI will become socially relevant if it is designed in an interdisciplinary, responsible and practical way.