By Mario J. Paredes
The emergence of Artificial Intelligence (AI) in healthcare is being framed as a dilemma between technological innovation and human care. Still, public opinion data show that society has already drawn a very precise boundary between these two worlds. This is confirmed by various studies conducted by Ipsos, covering 33 countries and totaling 24,132 interviews: the fifth edition of the 2026 AI Monitor (conducted among 23,532 adults in 32 countries, including a Spain-specific report); the international study, Global Campaign: Thank You, Doctor (based on 1,600 interviews across eight countries on four continents, conducted for the Pontifical Academy for Life and the Dr. Ramón Tallaj Foundation); and the analysis of Generation Z physicians. The overarching conclusion shared by citizens, patients, and young doctors is clear: AI is accepted as a tool that manages tasks and provides assistance under medical supervision but is rejected as an autonomous agent that diagnoses or makes decisions on its own.
The first finding of the reports is that AI already forms part of the healthcare routine for the digitally connected population. In eight countries analyzed in Thank you, Doctor (the United States, the United Arab Emirates, Italy, China, Mexico, Chile, South Africa, and the Dominican Republic), ninety percent of respondents have used AI; within this group, forty percent frequently consult it about health and 15% turn to it whenever they have a medical problem. In addition, respondents are familiar with an average of 3.9 wellbeing technologies, led by mobile health apps and smartwatches that measure vital signs (81% each), digital prescriptions and medical records (69%), and telemedicine (54%).
However, using technology does not imply mindlessly trusting it. The 2026 AI Monitor in Spain report summarizes this paradox under the slogan “I don’t trust it, but I use it”: sixty percent of Spaniards – and 63% of people worldwide – use AI tools every day, even if they do not always trust them. Sixty-two percent review the results that AI tools generate. Likewise, although 74% say they know what AI is, only 48% know how to identify which daily services use it, a figure that drops to 44% among women and 40% among those over 50.
In primary care consultations, citizens comfortably approve of AI taking on bureaucratic and preventive tasks, but they draw a red line when its role shifts to clinical judgment. Patients feel comfortable if their doctor uses AI to track medications (77%), manage appointments (68%), give preventive recommendations (61%), perform continuous monitoring (58%), perform triage in the emergency room (55%), or analyze prior tests (55%).
Conversely, comfort levels fall to 45% when AI is used to select treatments and 44% when it is used to diagnose illnesses. In Spain, when AI is presented as performing these tasks without mentioning physician oversight, discomfort reaches 51% for treatment selection and 50% for diagnosis, respectively. In fact, patients believe AI should never replace diagnosis (59%), analysis of previous tests (53%), or choosing treatments (48%). For the majority of respondents, the ideal role of AI is to help the doctor (39%) or the patient (37%), and when seeking a second medical opinion, 31% would use AI only as a guide, deferring to the doctor’s judgment, while another 31% would use it as a screening tool to determine whether to consult a second human specialist. Only 11% would trust the algorithm more than the doctor.
Trust in doctors continues to rest on human and scientific factors: clear explanations (47%), medical competence and up-to-date knowledge (44%), active listening (44%), honesty (41%), empathy (36%), and time spent during appointments (33%), leaving the use of advanced technology and AI in a last-place ranking, at 24%. While 28% fear AI will make medical consultations feel less personal, 27% believe it will improve empathy by freeing doctors from administrative tasks and letting them spend more time listening to patients.
In the international comparison, the dividing line is not whether people use AI, but how much autonomy they are willing to grant it. The most enthusiastic group includes China (0% rejection, 86% convinced that AI will democratize healthcare, and 75% willing to share their data), the United Arab Emirates (the leader in frequent AI use for health, at 50%), and Mexico (63% willing to share their medical records, although the country also leads the world in concerns about discriminatory bias, at 61%). Chile occupies a middle ground (the leader in telemedicine use, at 90%, and in demands for transparency, at 81%, but with 38% rejecting autonomous diagnosis), South Africa (where 25% consider AI just as credible as a medical specialist, alongside the highest level of concern about cybersecurity), and the Dominican Republic (which, alongside the United States, leads expectations for preventive medicine, at 60%, but where 42% insist that AI should serve only as a complement to the doctor). The most skeptical group is led by the United States, where 25% reject the use of AI in healthcare decisions, 49% oppose allowing it to select treatments, only 42% would share their data, and active listening (26%) far outweighs technology (4%) as a basis for trust. Italy is also among the most skeptical: only 33% believe AI will democratize healthcare, and just 6% accept the replacement of clinical tasks by AI.
The main global concerns about AI in healthcare are the lack of clarity about who bears legal responsibility when it makes a mistake (71%) and diagnostic errors (70%), followed by the dehumanization of care (63%) and data breaches (62%). As a result, society is demanding three key safeguards. The first is non-negotiable human medical oversight, with a human doctor validating every decision. This is demanded by 80% of respondents globally, rising to 89% in China, 85% in Mexico, 83% in the United States, and 81% in Italy. The second is explicit transparency whenever AI is used, a requirement supported by 82% of patients and 72% of respondents in Spain. The third is ethical accountability, which respondents believe should rest with medical associations (35%, rising to 53% in the United States) and governments (34%), ahead of technology companies (20%). Under these conditions, 55% would share their medical records to help prevent disease, and 56% believe AI will democratize healthcare. Looking five years ahead, respondents expect AI to play a role in personalized prevention (52%), routine medical conditions (50%), clinical assistance (46%), and administrative management (45%).
This aspiration for AI to take over administrative burdens and give doctors more time with patients resonates strongly with Generation Z physicians (born between 1997 and 2012). They no longer want simply to be “doctors”; they want to be people who practice medicine. They see their work as a social commitment to helping patients, one that is compatible with their own quality of life. They prioritize mental health, collaborative working relationships, scientific transparency without marketing spin, and tools that make consultations more efficient. Ultimately, patients and young doctors agree that algorithms should work in the background, preserving the integrity of clinical judgment, attentive listening, and the human dimension of medicine. This reminds us that SOMOS has already made significant strides in advancing this humanistic approach. Through its Thank You, Doctor campaign, SOMOS has sought to refocus attention on the direct, personal relationship between doctors and patients that has always been at the heart of medical care. The mission is no longer solely to protect and heal the body, but also to provide comprehensive care and support for the whole person.