Most Jordanian Pharmacists Welcome AI for Counseling, but Trust Gaps Remain

Artificial intelligence is quietly reshaping the way medicines are dispensed and explained, and one of the clearest tests of that transformation is now underway in community pharmacies. A new cross-sectional study from Jordan, published in BMC Health Services Research, offers one of the most detailed snapshots yet of how frontline pharmacists actually feel about handing parts of their counseling role to machine intelligence. The verdict is a striking mix of enthusiasm and unease: most pharmacists already use AI tools in their daily lives and more than half want to bring them into medication counseling, yet nearly three-quarters worry that regulation has not kept pace with the technology.

The research team, led by Manal S. Kaawsh of Applied Science Private University in Amman together with colleagues across Jordan, the United Arab Emirates, Qatar, Yemen, and Iraq, surveyed 300 community pharmacists between February and March 2026. Their instrument was built around the Technology Acceptance Model, or TAM, a framework developed in the late 1980s that predicts whether people will adopt a new technology based on two core beliefs: perceived usefulness, meaning the belief that the tool improves performance, and perceived ease of use, meaning the belief that operating it requires little effort. By measuring these constructs alongside behavioral intention, prior AI experience, perceived benefits, and specific concerns, the researchers could statistically model which psychological levers actually drive a pharmacist’s willingness to adopt AI in practice.

The baseline familiarity numbers are remarkable. Of the 300 pharmacists surveyed, 283, or 94.3 percent, said they were familiar with AI tools, and 222, or 74 percent, reported using them in everyday life. Yet formal education has clearly lagged behind personal experience: 185 pharmacists, or 61.7 percent, had never received any structured AI training. That gap between informal fluency and professional preparation is one of the study’s most consequential findings, because it suggests pharmacists are forming their attitudes toward clinical AI based on consumer tools like chatbots rather than validated, pharmacy-specific systems.

When asked directly about adoption, the cohort split in a way that captures the current moment in healthcare AI. Exactly 156 pharmacists, or 52 percent, said they were willing to integrate AI into medication counseling. Another 104, or 34.7 percent, were uncertain, and 40, or 13.3 percent, were unwilling. The large uncertain middle is arguably the most important group. These are practitioners who neither embrace nor reject the technology, and their eventual position will likely be determined by how well regulators, educators, and software developers address the specific concerns the survey catalogued.

Those concerns are concrete and technical. The most frequently cited was inadequate regulation, flagged by 218 pharmacists, or 72.7 percent. Close behind were doubts about AI accuracy and worries about data privacy, each raised by 205 respondents, or 68.3 percent. For medication counseling specifically, these fears are well founded. Large language models can generate plausible but incorrect drug information, a phenomenon known as hallucination, and counseling errors involving dosing, interactions, or contraindications can directly harm patients. Privacy is equally serious, since counseling often touches on sensitive diagnoses, and pharmacy systems may not meet the encryption and access-control standards that health data protection requires.

On the other side of the ledger, pharmacists saw substantial upside. Improved patient education was the top perceived benefit, named by 204 participants, or 68 percent. Drug interaction checking followed at 200 respondents, or 66.7 percent, and counseling efficiency at 197, or 65.7 percent. These benefits map neatly onto the real bottlenecks of community pharmacy. Interaction checking is a computationally intensive task where AI can cross-reference a patient’s full medication list against vast pharmacological databases in seconds. Patient education is a time-intensive task where AI-generated, language-tailored explanations could free pharmacists to focus on judgment and empathy rather than rote explanation.

To move beyond simple percentages, the team ran a logistic regression, a statistical technique that estimates how each factor independently changes the odds of a binary outcome, in this case willingness to adopt AI. The model produced a clear hierarchy of drivers. Behavioral intention emerged as the primary source of willingness, with an adjusted odds ratio of 1.966, a 95 percent confidence interval of 1.212 to 3.188, and a p-value of 0.006. In practical terms, pharmacists who already intend to use AI were nearly twice as likely to be willing adopters, which is consistent with the TAM’s central claim that intention is the strongest proximal predictor of behavior.

Age also mattered, and in the expected direction. Younger pharmacists were more willing to adopt AI, with an adjusted odds ratio of 0.947 per year of age, a 95 percent confidence interval of 0.915 to 0.980, and a p-value of 0.002. Because the odds ratio is below one, each additional year of age slightly reduces the odds of willingness. This generational gradient mirrors findings across digital health research and suggests that AI integration in pharmacy may follow a cohort-replacement pattern, with adoption accelerating as digitally native graduates enter the workforce. Notably, prior AI experience and perceived ease of use did not reach statistical significance in the final model, hinting that hands-on familiarity alone does not convert into professional adoption without intention and enabling conditions.

Perhaps the most reassuring qualitative signal in the data is how pharmacists framed AI’s role. The authors conclude that community pharmacists generally viewed AI as a valuable adjunct to medication counseling rather than a replacement for professional judgment. That framing matters for the future of the profession. If AI is positioned as a decision-support layer, handling information retrieval, interaction screening, and first-draft patient explanations, the pharmacist retains accountability for the final clinical encounter. This division of labor aligns with the concept of intelligence augmentation, in which machine capabilities extend rather than substitute for human expertise.

The path forward, according to the study, runs through four gates: regulation, training, data privacy, and reliability. Jordan’s findings are likely generalizable across many middle-income health systems where community pharmacies are the most accessible point of care and where AI tools arrive faster than the rules governing them. For developers, the message is that accuracy guarantees and privacy-by-design architecture are not optional features but adoption prerequisites. For pharmacy schools, the 61.7 percent training gap is a curriculum to-do list. And for regulators, the 72.7 percent who cited inadequate oversight are issuing a warning that trust, not technology, is now the limiting reagent in pharmacy’s AI future.

Subject of Research: Community pharmacists' acceptance and willingness to adopt artificial intelligence for medication counseling in Jordan

Article Title: Community pharmacists’ acceptance and willingness to adopt artificial intelligence for medication counseling: a cross-sectional study in Jordan

Article References: Kaawsh, M. S., Alzoubi, K. H., Barakat, M., Al-Ashwal, F. Y., & Abu-Farha, R. K. (2026). Community pharmacists’ acceptance and willingness to adopt artificial intelligence for medication counseling: a cross-sectional study in Jordan. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15820-4

Image Credits: AI Generated

DOI: 10.1186/s12913-026-15820-4

Keywords: artificial intelligence, community pharmacists, medication counseling, Technology Acceptance Model, Jordan, pharmacy practice, behavioral intention, data privacy, AI regulation, drug interaction checking, patient education, health services research