The Ministry of Health has introduced a new regulatory framework governing the use of artificial intelligence language tools and large language models in healthcare, allowing medical and administrative professionals to benefit from AI while placing clear limits on its use in clinical decision-making.
Under a decision issued by Undersecretary Sheikh Dr. Salman Al-Sabah, the Ministry adopted a practical guide covering the use of tools such as ChatGPT, Claude, Gemini and OpenEvidence in health, administrative, educational and research activities.
AI cannot replace doctors
The guidelines make clear that AI is an assistive tool, not a substitute for doctors or healthcare professionals. AI-generated information must not be treated as a final clinical reference or recommendation.
Healthcare professionals remain fully responsible for decisions made within their professional duties. AI cannot be used as the sole basis for:
- Final diagnosis or treatment decisions
- Prescribing medicines or determining drug dosages
- Interpreting medical tests
- Patient admission or discharge
- Referrals and clinical triage
- Assessing disease severity
- Making decisions concerning an individual patient
Strict protection of patient data
The Ministry has placed strong emphasis on patient privacy and data security. Patient-identifying, confidential or sensitive health information must not be entered, uploaded or copied into AI tools, cloud services or platforms that have not been officially approved.
The fact that an AI tool is widely known or publicly available does not make it automatically suitable for use with Ministry or patient data. Approved tools must comply with applicable data protection, cybersecurity, cloud computing, confidentiality and legal requirements.
Three levels of AI risk
The guide categorizes AI use into three risk levels:
- Low risk: General drafting, translation, summarization and preparation of non-confidential educational or administrative content, provided there is human review.
- Medium risk: Summarizing health guidelines, organizing differential diagnoses and reviewing general medication information, with sources verified.
- High risk: Final diagnosis, treatment selection, prescriptions, dosage decisions, clinical triage and decisions involving individual patients. AI must not be relied upon for these decisions.
Special rules for Arabic health content
The Ministry also highlighted the need for careful review of Arabic-language AI content. Linguistic accuracy does not necessarily mean medical accuracy.
Any AI-generated, translated or simplified Arabic health information intended for patients or the public must be reviewed and approved by a qualified healthcare professional to ensure clinical accuracy, translation quality, cultural appropriateness and suitability for the audience’s health-literacy level.
No AI health advice in the Ministry’s name without approval
Health facilities and departments are prohibited from launching or promoting AI tools that provide health advice, treatment information, screening or clinical guidance in the name of the Ministry of Health or its facilities unless the tool has undergone formal evaluation and accreditation.
Such evaluation must consider clinical safety, legal requirements, cybersecurity, privacy, technical performance and the monitoring of potential errors.
AI errors must be reported
Healthcare facilities are required to report incidents involving AI errors, patient privacy breaches or unsafe use, including unauthorized entry of patient information into unapproved platforms or the use of unreliable AI-generated information that could affect patient safety or quality of care.
Healthcare workers can use AI for:
The Ministry permits AI for tasks such as drafting, translation, summarization, preparing non-confidential educational and administrative materials, reviewing health guidelines and general medication information, organizing information, supporting learning and preparing Arabic patient content, provided the required human review and approval are carried out.
The key message of the new framework is clear: Kuwait’s Health Ministry is opening the door to AI as a healthcare support tool, but medical responsibility remains firmly with qualified human professionals.
