AMdEX: a level playing field through an open data exchange infrastructure
A level playing field is an important condition to accelerate (cross-sectoral) data sharing on a large scale. We share this point of view with our participant AMdEX. Its main goal is to set up a data exchange infrastructure that operates as an open public facility for all that want to exchange their data. During the upcoming community meeting on March 25th, Wouter Los, consultant and initial coordinator of AMdEX will elaborate with colleagues on AMdEX’s activities and use cases with regards to (cross-sectoral) data sharing. We spoke with him in advance.
An open market as an alternative for gated data communities
“A level playing field can only be created when organisations change their perception of data sharing.” says Wouter, who makes the comparison with the way our food chain has changed over the years. “In the very beginning, people were all self-resourcing farmers that grew their own crops and kept their own animals. Over time, for the benefit of us all, we started to share and exchange our goods at an open market. Which not only gave access to a wider offering of goods, but also resulted in competition, new ideas, new entrepreneurs, and a flourishing economy.” According to Wouter, the data economy has to make a similar shift. “The far majority of organisations still operate as ‘self-sufficient’ farmers, keeping their data – and the knowledge it represents – for themselves. A small group of organisations have grown ‘their farms’ into huge gated communities by buying other organisations, while starting their own markets. These are the big platforms such as Google, Amazon and Microsoft. The main issue here is that one can only participate at these gated community markets by complying with their rules which they had set up in their advantage. For the data economy to work for all of us, a new market needs to be created. An open market that comes with different rules and a new infrastructure.”
A level playing field can only be created when organisations change their perception of data sharing.
The Internet of Data: an open exchange infrastructure for data sharing
AMdEX aims to offer an open exchange infrastructure for data providers and data users to exchange data in a safe and controlled way. “For this infrastructure to succeed, a new layer has to be added on top of the internet as we know it today”, Wouter explains. “Today, it is easy to use the internet to send data from A to B in a safe way. However, it is extremely difficult to stay in control of what the receiver of this data is allowed do with it. This is something AMdEX wants to build in ‘by design’. Just as natural as we use the internet for digital connectivity, ‘the internet of data’ serves as a mechanism in the background to exchange data both in a safe and controlled way.”
The internet of data infrastructure is based on an automated process of digitally enforced data sharing contracts, built upon supporting hardware and software for securing trust and audit. Wouter: “While some organisations are willing to exchange data, others might only provide access to see their data without providing permission to use it. Other organisations might not even be willing to provide access to their data, but only allow an (for example AI) algorithm to draw insights from this data. One could even think of a situation in which company A that owns the algorithm that visits the data from company B, does not allow company B to know how the algorithm works. The infrastructure facilitates all these levels of access to data, ensuring that organisations are in control of their data with no need to distrust one another.”
Generic agreements for data sharing are key
According to Wouter, standardisation of agreements for data sharing is key for this infrastructure to work best. Because of this, he foresees an important role for the Data Sharing Coalition. “It is easier for computers to understand and act upon digital contracts when they meet known, widely accepted standards. This makes the activities of the Data Sharing Coalition regarding the generic agreements for data sharing so important. When widely implemented, these agreements make data sharing at large scale much easier.”
According to Wouter, just like the internet itself, the internet of data should become applicable across borders, ultimately worldwide. “Thirty years ago it was extremely difficult to digitally connect with someone across the globe, whereas nowadays we consider it totally normal. I expect the same to happen with regards to data sharing, where the entire world can exchange data using one distributed and federated infrastructure with clear rules, based on the same standards. However, for this infrastructure to succeed on a large scale, collaboration on a large scale is also needed.” Wouter emphasises. “We have organised several workshops in collaboration with the European Commission to bring together like-minded European initiatives, including the French IMT Teralab, which is also involved in the Gaia X project.”
Testing the infrastructure with AI use cases involving SAE ITC
The data exchange infrastructure that AMdEX is working on, is intended to facilitate a wide variety of use cases, including a global effort with SAE Industry Technologies Consortia® (SAE ITC) through its ExchangeWell™ program, which supports AI development. An affiliate of standards organisation SAE International®, SAE ITC programs bring industry stakeholders together to achieve common benefits, whilst providing subject matter expertise and an anti-trust umbrella. Being active for over seventy years in the automotive and aerospace industry, SAE ITC enables organisations to define and pilot a wide range of best practices. These include creating trusted implementations of standards where organisations need to collaborate to achieve their target outcomes. Directed by Lisa Spellman, the ExchangeWell program enables digital transformation and data-driven results through collaboration and smart data management, and includes a pillar targeting the challenges to make data accessible and usable for AI development in a trustworthy way.
Wouter: “Hereto ExchangeWell adopted a concept, developed by AMdEX, called a Digital Data Marketplace. As such, this effort intends to support a standards development effort from a SAE/EuroCAE working group on AI for Aviation Systems. The focus is on standards making ML algorithm development safety assessable and certifiable. For aviation authorities, the involved data management and its traceability is a key aspect for AI certification. A Digital Data Marketplace, governed by an Industry Consortia e.g., created by SAE ITC, is considered a way forward to enable AI development in safety critical applications. Here multiple parties, such as manufacturers, operators and maintenance organisations need to collaborate on data sharing rules to allow each party to provide parts of the required data.”
“The work of the Data Sharing Coalition provides a useful base”, says Wouter. “Leon Gommans, representing Air France KLM is leading this SAE ITC standards effort and together with the University of Amsterdam, AMdEX and input from the Data Sharing Coalition, brings these collaborative efforts together by organising use-case testing in the AMdEX Fieldlab applying the Digital Data Marketplace concepts for AI development. A Digital Data Marketplace prototype was developed and demonstrated as part of an NWO research project called DL4LD, together with technology partners Equinix, Dell, Nokia and Ciena. This project demonstrated how data sovereignty could be maintained whilst an algorithm, supporting predictive maintenance, could train in a federated way across three locations.”
On March 25th, AMdEX will host a presentation on their vision and activities with regards to (cross-sectoral) data sharing. Next to Wouter Los, who elaborates on AMdEX, Leon Gommans, Science Officer at Air France-KLM and Lisa Spellman, Executive Director of Exchangewell (SAE ITC) will provide more context on their data sharing activities. Do you want to attend this community meeting? Do not hesitate to contact us.