Medical chatbots and ethical disorientation: from Dr Google to Dr ChatGPT: the voice of artificial intelligence in the care journey

In recent years, access to online health information has undergone a radical and profound transformation: we have moved from the passive information model of search engines (“Dr. Google”) to active, conversational interaction with chatbots based on large language models (LLMs). This shift represents not only a technological evolution but also an anthropological change in how patients interpret their own experience of illness and build their relationship with their doctor. A qualitative comparative analysis was conducted on five chatbots (GPT-5, Gemini 2.5 Flash, Claude Sonnet 4.5, Grok 4 Veloce, and DeepSeek), presenting each with the same seven medical questions in separate chats on the same day, using free-access versions to simulate the average user’s experience. Findings highlight the risk of a progressive disintermediation of medical knowledge, accompanied by phenomena of conversational manipulation and an implicit redefinition of clinical authority. In this context, the patient arrives at the consultation no longer as a seeker of information, but as an already informed interlocutor, carrying expectations, beliefs, and very often a self-diagnosis. This dynamic gives rise to an ethical disorientation involving both parties: patients, who may be influenced by simulated relational interractions, and physicians, who face a perceived erosion of epistemic authority. In light of this, there is a need to rethink and reframe the therapeutic relationship as an encounter between human vulnerabilities, in which artificial intelligence may serve as a supportive tool, but cannot replace the relational, ethical, and responsible dimension of medical practice