Nearly half of Germans consult artificial intelligence (AI) before seeking care from a primary care physician, reflecting a broader global shift in how people access health information and engage with clinicians.

According to a survey by the German digital association Bitkom, 45% of respondents said they consulted AI chatbots, such as ChatGPT, Gemini, or Copilot, instead of first seeking medical advice when health concerns arose. About 71% viewed AI in medicine positively, and 27% viewed it very positively. More than half (55%) reported that they understood chatbot responses better than the results from a conventional Internet search.

Christina Raab, vice president of Bitkom, emphasized in a statement that the majority of Germans are open to the use of AI in healthcare and actively use AI.

Trust Paradox

Public acceptance of AI does not eliminate these concerns. About 71% of respondents said they feared the misuse of personal data, and 69% worried that AI could reduce human attention during treatment. Many participants also expressed concerns about potential medical errors attributed to AI. Only 9% of the participants reported no concerns.

Simultaneously, the findings highlight a paradox: Although patients increasingly trust AI-generated advice, they may perceive physicians who use AI as less competent, empathetic, and trustworthy.

Evidence suggests that disclosing the use of AI in clinical practice may reduce patients’ confidence. In a study published in JAMA Network Open, Moritz Reis, MSc, from the Institute of Psychology, University of Wuerzburg, Wuerzburg, Germany, and colleagues found that physicians associated with AI rated lower on empathy, trustworthiness, and competence.

The researchers presented the participants with advertisements for medical practices that either mentioned or did not mention the use of AI. The participants then rated the physicians depicted in the vignette.

Speaking with Medscape’s German edition, Reis said, “Afterward, we asked participants how competent and empathetic they rated the physician and how likely they would be to make an appointment with him. There was clearly a limitation: Physicians associated with AI are perceived more negatively in terms of empathy, trustworthiness, and competence.”

Role Redefined

Approximately 30% of respondents considered chatbot guidance valuable as a physician’s second opinion. According to a survey by Bitkom,16% reported that they disregarded a medical recommendation in favor of AI-generated advice.

Rising trust in “Dr AI” and declining trust in physicians who use AI are reshaping doctors’ roles in healthcare delivery. Physicians express concern that they may struggle to maintain authority during traditional bedside examinations.

The German Medical Association cautioned in a position paper on digitalization that “knowledge competition” could emerge between clinicians and patients. The dynamics, as the association wrote, could create pressure to justify clinical decisions and potentially restrict therapeutic options.

In a commentary in the Journal of Medical Systems, Florian Reis, MD, from the Institute of Medical Informatics at Charité-Universitätsmedizin Berlin, Berlin, Germany, argued that patients who arrive with AI-derived knowledge may contribute to the “democratization” of the physician-patient encounter.

Reis suggested that the central challenge lies less in the technology itself than in the psychological context surrounding psychological factors, such as a lack of transparency and ambiguity.

If patients do not understand how AI tools are integrated into care and if digital systems are experienced as impersonal substitutes for human judgment, trust may erode.

“When AI is thoughtfully integrated and transparently disclosed, it has the potential to reinforce the very trust it threatens to displace,” wrote the authors.

Reis further argued that clinical decision-making should evolve toward greater transparency and participation. When presenting a treatment plan, physicians can explicitly reference both their expertise and AI-based contributions.

“It is no longer enough to simply teach doctors how to use technology. We also need to teach them how to talk about it,” he said. Similarly, algorithms may not be understood as competitors but as amplifiers of clinical care.

This story was translated from Medscape’s German edition.

This article is part of an ongoing, year-long Medscape series examining the decline of public trust in medicine. More content is available here.