AI and robotic systems in healthcare

The healthcare sector is currently characterized by a wide range of challenges. The shortage of skilled workers, ^{1 , 2} ageing societies, ^{3 , 4} and the increasing prevalence of chronic and infectious diseases ⁵ place significant strain on healthcare systems worldwide. In response, AI-driven technologies and robotics are often envisioned as promising solutions to enhance efficiency, improve care practices, and alleviate the burden on healthcare professionals. However, while expectations for these technologies are high, their successful integration into healthcare practice remains complex and demands thorough investigation. ^{6 , 7 , 8}

Much of the research on robotics in healthcare has traditionally been conducted in controlled laboratory settings. ⁹ However, transferring these technologies into real-world environments presents significant challenges, from tackling technical integration, to deploying appropriation processes, ¹⁰ and navigating ethical considerations. ¹¹ This Special Issue of i-com is dedicated to exploring how AI and robotic systems can be effectively designed, implemented, and evaluated within healthcare contexts. By focusing on long-term studies, living lab environments, and participatory approaches, the contributions in this issue offer insights into the practical application and impact of AI and robotic technologies in healthcare.

The five articles featured in this Special Issue cover a diverse range of applications and methodological approaches. The first three contributions present design case studies that explore the implementation and appropriation of different robotic systems in healthcare practice. ¹² Besides the design case studies, the issue includes two contributions which offer conceptual and methodological perspectives on the use of robotic systems in care contexts and on medical narratives for sharing complex medical knowledge.

The contribution by Unbehaun et al. (2025) ¹³ focuses on the use of robots to support recovery in orthopedic rehabilitation facilities. Interviews were conducted with therapists and patients, and observations were made in the facilities, focusing on the telepresence robot Temi (version 2), which was equipped with various digital applications (e.g., features for balance exercises, calf stretches, and squats). This approach helped identify opportunities and challenges in rehabilitation. Research during the patients’ three-week rehabilitation shows that personalized activities add value to patients’ motivation and recovery. This added value comes, for example, from meaningful interaction and music-based exercises. This article demonstrates that digital applications integrated into robots may contribute to therapeutic outcomes. This is achieved through personalized exercises, maintenance of rehabilitation processes, direct feedback and data-driven progress tracking. The focus is on co-creating a more engaging and interactive environment to promote successful orthopedic rehabilitation for patients.

The article by Schulz et al. (2025) ¹⁴ deals with the use of robots in healthcare as well. The authors referred also to Temi (version 2), and investigated the experience of its embodiment. Robots are perceived differently than, for example, tablets, because they are experienced by humans as other bodies in face-to-face interactions. Their work focuses on the experience of this embodiment in video consultation from a human-computer interaction perspective. To this end, experiments were conducted with 18 participants who took on the role of patients and conducted a video consultation with a person who took on the role of a physician. This medically trained person performed a brief examination to simulate a real assessment. The Temi robot was compared with an HTC Nexus 9 tablet for this. The participants’ experience was evaluated using quantitative and qualitative methods. The results show no significant differences between the tablet and the Temi robot in the presence experience. However, it was found that participants preferred communication that resembled face-to-face interaction with other people. In this context, this suggests a preference for technical artifacts such as robots that are perceived as having a body.

The paper by Manavi et al. (2025) ¹⁵ focuses on the interaction between the robot Pepper and people with mild cognitive impairments. In a four-month design case study at a nursing home they investigated how Pepper mediates attitudes, practices, and social contexts. Two app prototypes were developed and integrated into Pepper and tested over 10 weeks by a social caregiver, the former manager and 10 residents. The focus was on a qualitative assessment of social interaction and group-based user experience, including video analysis. Their results show that Pepper can support everyday, individual and social activities and foster meaningful interactions. Residents initiated spontaneous conversations with Pepper, for example, one resident discussed family problems. This revealed a discrepancy between Pepper’s actual capabilities and the sensemaking that took place, which can lead to varied uses. The resident perceived Pepper’s movements (e.g., nodding, blinking) as active listening and responsive behavior. The study illustrates how individual appropriation enables diverse uses, offering insights into the design possibilities and appeal of Pepper in human-robot interaction. Manavi et al. (2025) ¹⁵ stress the need for ethical reflection. Fully delegating caregiving to robots requiring empathy is questionable. They recommend that robots should be accompanied by humans and clearly identified as machines to support responsible decision-making.

Lienenbrink et al. (2025) ¹⁶ provide a conceptual tool designed to assess the potential implementation of robots in care contexts. The assessment tool, entitled “READY?”, is designed to support stakeholders from care facilities and services in their reflection on the use of robotic systems for nursing care. It includes a digitally supported question catalog, collaboratively completed in on-site workshops with multiple stakeholder groups. The catalog covers the following categories: “Care”, “Privacy and Legal Issues”, “Ethical Criteria”, “Technology and Infrastructure”, “Institutional and Social Embeddedness”, and “Economic Criteria”. The “READY?” tool was developed in several iterative steps, including expert interviews, focus groups, workshops, and surveys. Ultimately, the tool was tested in care settings and evaluated through focus groups and a questionnaire survey. This showed that the tool can initiate debates about robotic systems for care in various organizations and strengthen collaboration in the technology implementation process among affected stakeholders in care contexts.

Mittenentzwei et al. (2025) ¹⁷ also address the implementation of digital technologies in healthcare, focusing on medical narratives. The paper presents two stories about people with non-alcoholic fatty liver disease. The first story is about a woman who was able to recover through increased exercise and weight loss. The second story is about a man who developed non-alcoholic fatty liver disease due to lack of exercise and weight gain. The study used a GenAI-assisted pipeline to develop stories with both characters to illustrate the risks and prevention of the liver disease. The aim of the stories was to present these in an understandable way to raise awareness of health-related lifestyle factors. The GenAI-assisted pipeline allowed the authors to visualize appropriate characters in a time-saving manner, reducing the need for artistic skills. To test the approach, a crowdsourcing case study was conducted with participants. The participants were asked to evaluate the consistency (over time) and authenticity (based on the data) of two designed characters. Furthermore, four experts qualitatively evaluated the stories with the two characters. They emphasized the importance of realistic representations of the body in order to recognize abdominal fat as a sign of non-alcoholic fatty liver disease. In this regard, photorealistic representations of the characters were considered more appropriate, emotionally appealing, and believable than comic versions, which were perceived as too abstract. The depiction of the context in the background of the images was judged differently. In one case, the man was seen in a fast food restaurant, which was perceived as too stereotypical. However, it was also noted that such visualizations help to emphasize lifestyle factors. The experts pointed out that the needs of the target group, the differentiation of emotions and the consistency of the representations should be in the foreground. The goal is to target people who belong to risk groups defined by aspects such as avoidable risks (e.g., nutrition) or age.

A recurring theme across all five contributions is the role of sense-making in the appropriation and evaluation of digital technologies. ¹⁸ How do healthcare practitioners perceive and reflect on these technologies? How do patients and caregivers interact with them? What methodological approaches are most effective in assessing their impact? These questions underscore the necessity of interdisciplinary and participatory research in this field also in the future.

Finally, the studies were conducted in real-world healthcare settings, moving beyond theoretical discussions to address the practical realities of AI and robotics in healthcare, thus emphasizing new insights into the socio-technical dimensions of healthcare innovation. With this Special Issue, we aim to contribute to further research and discussion on the integration of AI and robotics in healthcare, ultimately contributing to more practice-based, ethical, and user-centered solutions.