Humanoid robots already impress with communication and understanding, but they do not yet have truly skilled "hands". A lecture from the Department of Applied Informatics at Matfyze in Bratislava showed where the research stands, what the prospects are in healthcare, and why trustworthiness is crucial. We offer an overview of possibilities, limits, and recommendations for their sensible deployment.
Where we are today: strengths and weaknesses
The greatest progress has been in communication: robots can conduct a conversation, understand the scene, and even read emotions quite well. Manipulation, however, is weaker – fine motor skills, tactile perception, and proprioception lag significantly behind. Today they can capture attention, gather and evaluate information, or monitor situations, but they are not suitable for handling a patient, performing surgery, or fully replacing a nurse or caregiver. These skills will come later, on a timescale of several years – certainly not immediately.
Healthcare: benefits, risks, and how to get started
In practice, communication tasks in particular have proven effective: for example, making pediatric patients' stay more pleasant, reducing stress, delivering information, or collecting questionnaires. A humanoid's advantage is its adaptation to the human environment – it can open doors, call an elevator, or handle simple local manipulations. The robot is also easy to disinfect and more resistant to harmful influences; in crisis situations it can "vent" aggression that would otherwise be directed at the staff. For many heavy or routine transfers, however, specialized, non-humanoid robots are better suited today.
Deployment brings risks: unreliability in a busy real-world environment, poor connectivity, unexpected situations, and the limits of learning from examples. Data protection and cybersecurity are crucial, as is acceptance by staff and patients, including clear ethical "guardrails". A pilot with a narrowly defined set of activities and clear success metrics is recommended. The biggest mistake is an ill-defined or overly broad ambition – a well-targeted pilot delivers quick, measurable results.