The future role of artificial intelligence (AI) was the focus of the parliamentary evening on 26 February 2025, to which the L3S Research Center invited members of the Lower Saxony state parliament to the neighbouring Leibnizhaus. After a long day of plenary session, many members of parliament took the opportunity to find out more about the challenges and opportunities of artificial intelligence as well as current L3S research and transfer projects. The L3S is one of the most important AI research institutions in Germany and, with 200 researchers, the largest AI centre in Lower Saxony.
In her welcome address, Barbara Otte-Kinast, Deputy President of the Lower Saxony State Parliament, acknowledged the role of the L3S for AI research in Lower Saxony. Falko Mohrs, Lower Saxony’s Minister for Science and Culture, also emphasised the importance of research for innovation and prosperity in the state. The state of Lower Saxony supports the L3S Research Center with annual basic funding of €2 million. In addition, the L3S raises around €13 million in public and private third-party funds each year.
It is already clear that AI is having an impact on society, business, education and government, whether through the introduction and application of AI methods or through discussions about the opportunities and risks of the technology. The presentations by L3S directors Professor Dr Wolfgang Nejdl and Professor Dr Bodo Rosenhahn showed how modern AI approaches are making groundbreaking progress both in personalised medicine and in the development of comprehensible, trustworthy systems.
Artificial Intelligence for Personalised Medicine
In his talk ‘Dr AI – can it help yet?’, Prof Nejdl explained how deep learning is transforming medical diagnostics. He presented the historical development from traditional diagnostics to personalised medicine, using examples such as ImageNet and AlexNet. Particularly impressive was the presentation of advances in medical image analysis, such as tumour segmentation using nnU-Net, and the use of large language models (GPT-3, ChatGPT, GPT-4) to answer complex medical questions. These innovative methods enable more precise and personalised treatment planning and open up new perspectives in medical diagnostics.
Trust in deep neural networks
In his presentation ‘Talk to me! How to gain trust in a deep neural network’, Prof Rosenhahn addressed the challenges of explaining deep learning models. He emphasised that conventional neural networks are often seen as opaque “black boxes” and argued for approaches that make their decision-making processes transparent. Methods such as attention and saliency maps, as well as new concepts such as self-explaining neural networks (SENN) and Q-SENN, can make the underlying semantic concepts visible and thus increase trust in AI systems.
This hybrid intelligence approach – the interplay of human expertise and machine power – is a key building block for the future development of trustworthy AI applications.
Questions and answers
Both presentations made clear that the combination of innovative AI processes and understandable explanatory approaches is crucial for realising the full potential of artificial intelligence in healthcare and other applications. In the discussion moderated by L3S member Prof Dr Marius Lindauer and Dr Johannes Winter, Chief Information Officer of L3S, participants had the opportunity to discuss their questions about AI with L3S experts, including L3S member Prof Dr Margrit Seckelmann on legal issues and Prof Dr Thomas Seel on the future of AI and robotics in industry.
The more than 70 participants chatted over finger food and drinks. 15 posters and three robots provided insights into current research and transfer projects at L3S, such as CAIMed, the Lower Saxony Centre for Artificial Intelligence and Causal Methods in Medicine, and the European Digital Innovation Hub for AI and Cybersecurity (DAISEC). The guests took the opportunity to discuss AI topics with the scientists.
The evening highlighted the L3S Research Center’s focus on forward-looking, interdisciplinary AI research that combines technological progress with social responsibility.
Photos: Patrick Pötsch
