At Al-Hamshari Hospital in Sidon, a generative clinical assistant is piloting an innovative form of emergency healthcare support.

Talking about artificial intelligence When applied to healthcare, thoughts almost always turn to large university hospitals, research centers, or the highly digitalized facilities of the most advanced economies. Yet one of the most significant trials in recent months is taking place neither in Europe nor the United States, but rather within a hospital that operates daily in one of the most challenging humanitarian settings in the Middle East. Here, the goal is not to increase the efficiency of an already advanced system, but to ensure continuity of care in a context where electricity, connectivity, and qualified personnel are limited resources.

The project concerns theAl-Hamshari Hospital, located near the Palestinian refugee camp of Ein El-Hilweh, on the outskirts of Sidon, in southern Lebanon. The facility is managed by the Palestine Red Crescent Society, a healthcare organization that has been providing assistance to Palestinian refugees in the country for decades. This hospital launched the first structured program testing a generative clinical assistant specifically developed to operate in environments characterized by humanitarian crises, infrastructural instability, and severe pressure on healthcare services.

The experiment was developed by the startup Rhazes AI, a healthtech company with operations in UK and Qatar, created with the goal of using generative language models to reduce the administrative burden on doctors and provide decision-making support during clinical visits. Unlike many systems designed for large, fully computerized hospitals, Rhazes was developed to function even where computers, fast connections, and IT infrastructure are virtually nonexistent.

The choice of Lebanon was not accidental. For the developers, the value of the experimentation consisted precisely in verifying whether a tool clinical AI was truly capable of operating under the most challenging conditions: intermittent power supply, unstable internet coverage, limited availability of IT devices, staff forced to manage extremely large numbers of patients and the need to make clinical decisions very quickly.

Every month the hospital follows more than 4.000 patients thanks to the work of just 56 doctors e 31 nursesThese numbers reflect the constant pressure on healthcare teams. Many professionals are called upon to cover multiple disciplines simultaneously, from pediatrics to internal medicine, from surgery to infectious diseases, often without the immediate availability of a specialist.

Secondo Zaid Al-Fagih, co-founder and CEO of Rhazes AI, it is precisely in these contexts that artificial intelligence can make the greatest impact.

"Artificial intelligence tools shouldn't be confined to the healthcare systems of more developed countries. They should be introduced first where the need is greatest."

Before founding the company, Al-Fagih had worked in the National Health Service He was a British doctor and later served as a medical volunteer during the Syrian civil war. It was precisely that field experience that contributed to the project's birth: observing professionals forced to make extremely complex decisions without the ability to consult with specialist colleagues showed him how the immediate availability of reliable clinical information can directly impact the quality of care.

An essential health facility for tens of thousands of refugees

The Al-Hamshari Hospital It is one of the main healthcare facilities for the Palestinian camps in southern Lebanon. In addition to emergency, internal medicine, surgery, obstetrics, and pediatrics departments, the facility hosts one of the few dialysis services available to the area's refugee population. During the most acute phases of regional crises, the number of surgeries and hospitalizations increases significantly, further straining already limited resources.

La Palestine Red Crescent Society manages a network of hospitals, health centres, clinics and emergency services in the country which complement the assistance guaranteed by theUNRWA, the United Nations agency dedicated to Palestinian refugees. For many families, these facilities represent the only option for receiving specialized care without facing unsustainable costs within Lebanon's private healthcare system.

Lebanon's economic situation makes the situation even more delicate. In recent years, the country has experienced one of the most severe financial crises in its modern history, accompanied by a sharp devaluation of the currency, energy supply difficulties, and growing problems accessing medicines. Added to these challenges are the effects of military tensions along the southern border and the progressive increase in demand for healthcare.

In this scenario, even the simple availability of IT tools represents a luxury. As Al-Fagih explained, a pediatric department of a modern hospital in Dubai can have dozens of computers, while Al-Hamshari has only one shared terminal for each floor of the building. For this reason, Rhazes AI was designed to operate directly on smartphone, accompanied by small portable printers that allow you to produce prescriptions, instructions and clinical documentation directly at the patient's bedside.

From smartphone to medical record in just a few minutes

The heart of the platform is a generative assistant capable of monitoring the entire medical visit. During the interview, it listens to the conversation, automatically transcribes relevant information, organizes the medical record according to a standardized structure, and suggests possible diagnostic pathways based on available scientific literature. At the end of the meeting, it prepares discharge letters, patient instructions, administrative codes, and documentation necessary for clinical management.

The design focused primarily on operational resilience. The software continues to function even in the absence of an internet connection, storing data locally and synchronizing it only when the network is available again. This seemingly secondary feature is crucial in hospitals frequently powered by diesel generators and subject to power outages that can interrupt access to online services for several hours.

One of the most interesting aspects of the project, however, concerns the role attributed to artificial intelligence. Rhazes was not conceived as a system designed to formulate autonomous diagnoses nor as a replacement for healthcare professionals. The platform is designed as a clinical assistant, always leaving the doctor with final responsibility for all therapeutic decisions. AI suggests possible differential diagnoses, retrieves updated guidelines, organizes the information collected during the visit, and automates much of the administrative documentation, but does not directly intervene in the choice of treatment.

This approach reflects an increasingly widespread trend in major Western healthcare systems. Over the last two years, numerous companies have developed tools AI medical scribe, capable of listening to the dialogue between doctor and patient and automatically transforming it into structured medical records. In the United States, such applications are being tested by large hospital groups with the aim of reducing the so-called administrative burnout, that is, the time taken away from direct care by filling out documents and reports. In Al-Hamshari's case, however, the stakes are even higher: it's not simply a matter of increasing efficiency, but of enabling adequate healthcare even when available resources are extremely limited.

For many doctors working at the Lebanese facility, the platform also provides ongoing training support. Those working in a hospital like this are often called upon to manage cases from very different disciplines. A professional may find themselves seeing a child with suspected meningitis, a chronic patient with kidney failure, and a person suffering from a trauma emergency in the same day. Having access to up-to-date clinical protocols and evidence-based references within seconds reduces the risk of errors and helps standardize the quality of care.

Imagine having to treat a child with suspected meningitis without a pediatrician to consult. Rhazes can provide the same clinical support a specialist would: immediately, in the doctor's language, and directly on their smartphone.

has explained Zaid Al-Fagih

In Sidon, results, data security and limitations still to be verified

The experiment, which began inAugust 2025, will continue until the end of the observation period established by the researchers. The analysis not only assesses operator satisfaction, but also measures concrete indicators such as the time required to complete clinical documentation, the accuracy of recordings, daily use of the platform, and the impact on patient flow within emergency departments and outpatient clinics.

Initial results show high levels of acceptance among healthcare professionals, who use Rhazes daily, especially for generating medical records and quickly accessing diagnostic information. Although final results are not yet available, the developers believe that reducing time spent on paperwork could allow doctors to increase the time they actually spend with patients.

Alongside the potential benefits, questions naturally also emerge which now accompany all applications of theartificial intelligence in medicineHealthcare data management is one of the most sensitive aspects. In contexts characterized by political and infrastructural instability, the loss or damage of paper medical records is a recurring problem. According to Rhazes AI, the transition to digital systems protected by authentication and encryption improves information security compared to traditional archives, provided adequate IT protection measures are adopted.

"Rhazes is an assistant, not a substitute. The doctor always makes the final decision."

Al-Fagih stressed

The ability to function as well offline It also represents a subtle but crucial feature. In crisis situations, connectivity can suddenly be interrupted due to power outages, infrastructure damage, or network problems. The system continues to record data locally, synchronizing it only when the connection is restored. This way, clinical activity is not dependent on constant internet availability, a requirement still essential for many applications based on large language models.

Experiments reveal a possible model for humanitarian healthcare

Rhazes AI's ambition goes beyond single hospitals. The company plans to adapt the platform to other contexts characterized by conflict, instability, or fragile healthcare systems, including Syria, Yemen e SudanEach implementation will inevitably require changes related to language, local healthcare organization, device availability, and operating conditions, but the principle remains the same: build a minimum level of clinical support even where technological infrastructure is virtually nonexistent.

This perspective also opens a broader reflection on the role ofhumanitarian healthtech, a rapidly growing sector aiming to bring advanced digital medicine tools to emergency settings. While in recent years, most investments have focused on highly digitalized hospitals in the Gulf, Europe, and North America, initiatives like Al-Hamshari's demonstrate how the most transformative applications can emerge precisely in places where resources are scarcest.

The challenge, however, isn't just about the quality of algorithms. Healthcare professionals' training, the economic sustainability of platforms, the availability of minimal infrastructure, and the definition of shared rules on clinical responsibility and personal data protection will also be crucial. Without these elements, artificial intelligence risks adding complexity rather than reducing it.

The experience of Al-Hamshari Hospital suggests a different approach. Here, innovation is not measured by the number of servers installed or the available computing power, but by the ability to allow a doctor to dedicate a few extra minutes to the patient in front of them. In a hospital powered by diesel generators, where every visit often represents a response to long-unmet healthcare needs, even a simple smartphone can become a tool capable of increasing the continuity of care. If the final results of the trial confirm the initial indications gathered in the field, the project could help redefine not only the way artificial intelligence is used in clinical practice, but also the role it could play in healthcare systems operating in the most vulnerable areas of the world.

Here are three insights that might interest you:

A cutting-edge healthcare hub based entirely on AI in Geneva
The Silent Revolution in Lapland's Cottage Hospitals
North Macedonia and Slovakia: Digital Healthcare Alliance