What expectations do patients in the German healthcare sector have of AI-based chatbots? The results of the study conducted by Luisa de Alzaga Achter under the direction of Prof. Dr. Klaus Mühlbäck and Daniel Khafif, both lecturers at the ISM Campus in Munich, show that they are not very high. She used guided interviews to ask eight experts from different age and professional groups about their expectations, motivations and barriers to using chatbots. The interviewees came from different areas: Chatbot developers, healthcare representatives and patients, i.e. users of these chatbots.
Is the AI easy to use?
First of all, they hope for guidance and support from the chatbots. They want to be guided safely through the complex healthcare system with the help of technology in order to confirm their own actions.
The usability of the bot plays a major role here. Functionality, quality of responses, machine intelligence and user-friendliness are key requirements. “Patients understandably don't want to retell their medical history over and over again,” explains Luisa de Alzaga Achter. “They therefore expect the artificial intelligence to learn and remember the data they have already entered.”
Is my data protected?
The question of data protection quickly arises. Especially when it comes to sensitive information, such as their own medical history, respondents attach great importance to the responsible handling of their data by the AI. One option would be for the chatbot to be awarded a certificate that guarantees that data is stored securely and is easy to track. Patients also believe that data protection is best achieved if the chatbot is operated by a German company so that the data is also managed in accordance with German law.
Can an AI be human?
Another point of criticism from respondents is the characteristics in which humans are still far ahead of artificial intelligence: Language and empathy. Chatbots do not yet have complete language skills on a human level, which leads to misunderstandings and dissatisfaction among users. Artificial intelligence is also unable to match doctors when it comes to the empathy required in dialog with patients. Doctors also have reservations about AI-based chatbots. They like to get an overall personal impression in order to make a well-founded diagnosis.
How will AI be used in medicine in the future?
According to the current state of technology, AI only has limited applications in medical diagnostics and cannot replace a visit to the doctor. Lecturer Daniel Khafif summarizes: “It remains to be said that diagnostic AI systems should only ever be seen as assistance tools. The final decision-making authority for prescribing and dispensing medical treatment must remain with the human being, i.e. the doctor or nurse.”
However, AI could usefully complement medical professionals, as Luisa de Alzaga Achter has worked out. In particular, chatbots can provide valuable services for organizational tasks relating to finding and booking appointments as well as in prevention and aftercare. They can also collect more data and diagnoses than a single doctor and can therefore produce more comprehensive analyses and reports on illnesses. This, in turn, is in the interests of health insurance companies, as it allows them to offer cost-optimized and more efficient treatment options. These technologies significantly relieve the burden on medical staff by allowing AI to take over time-consuming communication tasks,” emphasizes Khafif. In times of staff shortages and cost pressure, the resources saved can be invested more efficiently in clinics and practices, which benefits both medical staff and patients.”
“Nevertheless, there are still some obstacles to overcome,” adds Luisa de Alzaga Achter. “A high level of user-friendliness, reliable functionality and security of data storage and processing combined with patient-friendly language and elements of human empathy could strengthen patient trust in AI-based chatbots.”
