‘Our challenge is to de-risk the use of AI in Dutch industry’

The ELSA Lab for Technical Industry recently received €2.4 million from the Dutch Research Council (NWO). Three researchers from the University of Groningen involved in the project share their plans for the ELSA Lab. ‘In industry, it is less complicated and safer to showcase the potential of AI.’

The project focuses on the ethical, legal, social and economic (ELSA) challenges of AI adoption in the technical industry. The uptake of AI has been relatively slow in Dutch industry, especially among SMEs, which could hurt global competitiveness. ‘Whether the Netherlands is behind depends on your frame of reference,’ says Ming Cao, lead applicant and full professor of Networks and Robotics at the University of Groningen. ‘But compared to the United States and China, Europe is much behind on AI adoption in industry.’

Cao (also scientific director of the Jantina Tammes School of Digital Society, Technology and AI) mentions large-scale warehouses as an example. US companies such as Amazon have already largely automated operations, partly with AI. ‘Warehouse robots are a fast-evolving area in AI applications for robots, but the Netherlands is lagging behind in this field.’ This also applies in other areas, such as assembly and manufacturing.

Opportunities and uncertainties

Co-applicant Christos Emmanouilidis, associate professor at the Faculty of Economics and Business at the University of Groningen, observes a similar development. Still, he sees opportunities for the Netherlands and Europe. ‘The original breakthroughs in AI came from the US, but the barriers to AI applications are getting lower. There are also an increasing number of success stories from Europe. This is good news given global uncertainties. We need autonomy on our continent.'

The project therefore focuses on the opportunities for the technical industry in the Netherlands which align also with public values and human rights. ‘Most industry partners have already started pilot projects to investigate what is possible with these emerging technologies.’ The challenge lies in the uncertainties regarding security, regulation and legislation. For example, what about intellectual property: how can companies use ‘large language models’ without disclosing sensitive information? The challenge is to make the use of AI less of a risk for companies. ‘We want to use the technology, but exclude the negative aspects as much as possible.'

Cao also points to legislation in Europe. Since 2024, the AI Act has been in place, establishing frameworks for the deployment of AI, for example regarding issues such as privacy. Starting from the ELSA approach, the AI Act has a central role in the project, the professor argues. 'On the one hand, the AI Act gives direction, but it also brings uncertainties, for example in the field of camera usage. Robots do not have eyes or ears: they use cameras to gather information about their surroundings. The AI Act contains strict rules on the use of cameras. This could slow down the adoption of AI in industry. After all, how do we ensure that robot cameras do not violate workers' privacy?'

The project takes these aspects of legal uncertainty very seriously. This is also reflected in the choice to host the PhD candidate funded from the project at the Faculty of Law. As Jeanne Mifsud Bonnici, full professor of Technology Law and Human Rights points out ‘Our faculty has a strong community of researchers on the role of law in AI developments.’ The research in the ELSA Lab for Technical Industry (ELSA4TI) complements research that is ongoing at the Faculty of Law on AI and workers’ rights and AI in healthcare (as part of the work in the ELSA AI lab Northern Netherlands coordinated by associate professor Mirjam Plantinga, at the UMCG).

Use cases

To test the ethical, legal, social and economic challenges in a real-world setting, the ELSA Lab project relies on ‘use cases’. ‘We do this in close collaboration with industry partners, such as Philips,’ Emmanouilidis explains. 'We aim to identify both the technological possibilities and potential risks. This not only benefits the industry, but also supports policymakers who ultimately define the regulatory frameworks. There is debate about what is the best way to regulate AI. The US takes a laissez-faire approach, while the EU, with the AI Act, adheres to stricter legislation based on people's fundamental rights. China, on the other hand, takes a different approach, with heavier government control. With the use cases, we are testing both the possibilities as well as the boundaries, to identify which approach works best for the industry.'

Unique to the project are the ‘use case families’, a cluster of industry partners with the same questions regarding AI applications, says Cao. This way, AI innovations can be applied more widely, as the challenges often apply to entire industries or sectors. Co-creation is also very important. The project involves very diverse partners, such as the University of Groningen, University of Twente and Tilburg University and 13 partners from industry. ‘We want to involve them in the project, and for this purpose we will apply the experience of co-creation in previous research projects to the ELSA labs’, Emmanouilidis states.

Collaboration

The Jantina Tammes School of Digital Society, Technology and AI can play an important role in this collaboration, says Cao. ‘It is a multidisciplinary subject with different stakeholders and a great societal impact, which fits well with the UG's strategy to become a fifth-generation university that connects with society.’ Above all, he hopes the project will address concerns people have about the adoption of AI. 'There is an important distinction between different domains. Health-related applications, for example, require different conditions than industrial ones, where the applications are likely to be less complicated. We have a long tradition with industrial robots: we know that robots can greatly improve productivity. In industry, it is less complicated and safer to showcase the power and potential of AI.'

The ELSA Lab project will be part of the National Growth Fund AiNed, executed by the AI Coalition 4 NL. More information on the Elsa Lab for Technical Industry, as well as the other Labs, can be found here.