Can technology and AI make everyday life easier for older people?

In the current episode of the Digital Public Health Podcast, host Rasmus Cloes speaks with Dr. Serge Autexier, head of the Bremen Ambient Assisted Living Lab (BAALL) at the German Research Center for Artificial Intelligence (DFKI) in Bremen. In a reconstructed apartment, the team researches how artificial intelligence in healthcare and assistive systems can support people in their daily lives.

Thumbnail showing Serge Autexier and the Podcast title.

Serge explains why the Living Lab does not showcase “smart home toys,” but instead tests real-world usability: Which technologies truly help? What is intuitive to use? And what does it take for an environment to become “intelligent” without overwhelming people or making decisions for them that they would rather make themselves?

A central topic is the gap between research and products. At BAALL, demonstrators are developed that often immediately convince visitors—accompanied by the typical question: “Why isn’t this available to buy?” Serge describes why project structures, responsibilities, and market-readiness often prevent good ideas from becoming products right away, even though individual components may later reappear in other contexts.

One concrete example is fall prevention with a “smart walker.” Practitioners noted that incorrectly used walkers are associated with a higher risk of falling. In one project, movement data was recorded, evaluated by physiotherapists, and used to train an AI that can detect incorrect use and provide guidance. This shows how closely AI in healthcare can relate to real prevention—and how quickly questions arise about medical device regulations, evidence, and funding.
Another topic is data integration: How can hospital data and everyday-life data be combined to improve predictions—for example, to increase quality of life for people after cancer therapy? Serge makes clear why this is so difficult in practice: data is often unavailable, ethical and IT security questions take time, and many projects ultimately leave too little room for proper evaluation.

Finally, Serge puts the AI hype into perspective. Large language models are not “simply part of the solution” in BAALL—data protection, local processing, and reliability are crucial, especially for vulnerable groups. The focus is therefore on robust, secure systems that can run on small devices, do not need to “listen in,” and won’t suddenly fail in everyday situations. He also explains why many assistive ideas fail not because of the technology but because of implementation pathways: Who will build it? Who is liable? Who pays for it? And who proves its benefits?

Watch/listen now and gain exciting insights: YouTube, Spotify, Apple, Dezer

More information: Website, Instagram

Further reading on the Topic
Christian Mandel, Jan Janssen, Maria Angelica Lopez-Ardila, and Serge Autexier. Smart Wheelchair with Semi- and Fully Autonomous Navigation, Journal on Technology and Persons with Disabilities Journal track of the 41st CSUN Assistive Technology Conference (ATC-2026), March, accepted, forthcoming, 2026

Serge Autexier. On Using Large Language Models Pre-trained on Digital Twins as Oracles to Foster the Use of Formal Methods in Practice, In Tiziana Margaria and Bernhard Steffen (Ed) 12th International Symposium Leveraging Applications of Formal Methods, Verification and Validation. Software Engineering Methodologies (ISOLA 2024),Crete, Greece, October 27–31, 2024, Part IV, Vol. 15222, LNCS, Springer Cham, October, 2024 link.springer.com/chapter/10.1007/978-3-031-75387-9_3

Serge Autexier, Christoph Lüth, and Rolf Drechsler. In M. A. Pfannstiel(Ed) Künstliche Intelligenz im Gesundheitswesen, Chapter Das Bremen Ambient Assisted Living Lab und darüber hinaus – Intelligente Umgebungen, smarte Services und Künstliche Intelligenz in der Medizin für den Menschen, Springer Fachmedien Verlag Wiesbaden, March, 2022 link.springer.com/chapter/10.1007/978-3-658-33597-7_40