From Waterfall to Sandbox: Digitalisation Leads to New Innovation Methods
The much-discussed report by former European Bank President Mario Draghi on the competitiveness of the European economy highlights digitalisation as a significant opportunity for innovation. The report analyses that the lack of a strong European ICT sector is a major cause of the 'innovation lag' of the European Union compared to the United States. Draghi sees the development of European artificial intelligence (AI) as a way to catch up, particularly by applying generative AI in the industry to enable a more innovative manufacturing sector. It is taken for granted that this technology is based on European values.
The development of responsible digital technology, products, and services based on European values has been a common goal, even before the Draghi report appeared. For instance, in 2019, the European Commission established a High-Level Interest Group for AI, which set conditions for responsible AI, including transparency, safety, fairness, and human oversight. With the rise of generative AI, these conditions have only become more important and are embedded in the European AI Regulation that recently came into effect. Many organisations are now searching for ways to develop or use such responsible AI systems based on European values.
Organisations are seeking solutions in applying 'value-sensitive design'. In this design method, specific values such as transparency or fairness guide the development and application of (digital) technology. Previously, this was typically developed step-by-step through the 'waterfall method', where design criteria were established and elaborated. In value-sensitive design, the underlying values that the technology must meet are identified first. One way to ensure this is to bring all stakeholders together and have them determine the requirements for the technology. To establish whether the technology indeed meets these values, an experiment is conducted in a controlled environment (the 'sandbox') before application.
The European AI Regulation, which aims to ensure that AI development is based on European values, has adopted several principles of value-sensitive design. An important one is that (high-risk) AI systems must be tested in a 'regulatory sandbox' before use. This is done to determine whether a system complies with European rules and what (potential) consequences the system has for users, consumers, and citizens. This way, it is determined which AI systems are based on European values and thus allowed on the European market. In this manner, digitalisation not only leads to innovation within companies or organisations but also to innovation in the way digital systems are developed.
This reflection is based on the column by Anne Fleur in iBestuur (in Dutch) on 24 January 2025.