Artificial intelligence (AI)-based technologies are increasingly integrated into the healthcare system and are beginning to play a more active role in clinical decision-making. As a result, healthcare organizations and regulators are encountering a growing number of requests for clinical trials where AI-based tools may directly impact clinical practice.
This document is developed with the aim of supporting the responsible evaluation of such research, reflecting the core values of patient safety, privacy protection, transparency, accountability, fairness, and appropriate human oversight, while simultaneously enabling innovation that can improve the quality of medical care.
The document presents guiding principles for evaluating prospective interventional clinical trials that include AI-based tools and is intended to assist Helsinki Committees in conducting ethical, scientific, and risk-aware assessments of these studies.
Development of the Document
The principles were formulated through a multidisciplinary process involving experts from healthcare organizations, academia, ethics, technology, and regulation. The work was based on international guidelines and standards, including SPIRIT-AI, as well as practical experience gained from real-world case studies and ongoing dialogue with local Helsinki Committees.
This document is an updated version of previous guidelines and reflects accumulated experience, developments in ethical discourse, and changes in international regulatory perception.
Purpose and Objectives of the Document
The central purpose of this document is to provide Helsinki Committees with a structured and practical framework for reviewing requests for interventional clinical trials that include AI-based tools.
The document does not define when an interventional study is required. Instead, it focuses on the key issues to be examined once such a study is proposed, in order to support informed, consistent, and proportionate decision-making.
Target Audience
The document is intended for:
- Helsinki Committees reviewing AI-based research
- Researchers and physicians leading clinical research
- Healthcare organizations conducting or hosting AI-based trials
- Developers and sponsors of AI-based medical technologies
- Regulators and decision-makers involved in the oversight of clinical research
Structure and Key Principles
The document presents 16 guiding principles, formulated as key questions intended to support professional judgment, rather than as a mandatory checklist. The principles are organized into four main categories, reflecting the logical sequence of evaluating an AI-based intervention throughout the research process:
- Description of the AI System: Tool characteristics, intended use, model version, training process, regulatory status, prior evidence, and nature of outputs.
- Data Governance: Data collection and processing, storage, cybersecurity, privacy protection, bias mitigation, and integration into clinical systems.
- Study Design: Inclusion and exclusion criteria, performance evaluation, error identification, model updates, and criteria for early termination of the study.
- Disclosure: Representation of AI use in the research protocol, informed consent process, and user training.
Each principle is accompanied by clarifications and supplementary explanations, providing practical guidance to support consistent and in-depth application across a variety of research contexts.
An Evolving Document
Given the rapid pace of technological development, changes in ethical considerations, and ongoing progress in international regulatory frameworks, this document is intended to be a living document and will be updated according to insights and experience gained in the future.
Questions, comments, and feedback can be directed to: samd@moh.gov.il