English
Arabic
Chinese (Simplified)
Dutch
Esperanto
French
German
Greek
Italian
Japanese
Korean
Portuguese
Russian
Spanish
UrduInternational scientific association, GIN, publishes principles for using AI to develop clinical guidelines
Abstract: https://www.acpjournals.org/doi/10.7326/ANNALS-24-02338
URL goes live when the embargo lifts
The Guidelines International Network (GIN) has developed principles for the use of artificial intelligence (AI) tools or processes when developing clinical guidelines. A GIN working group conducted a scoping review and analysis of practical examples inform the recommendations. To the authors’ knowledge, this is the first time the use of AI in health guidelines has been addressed in a framework for AI in health care. A summary of the development process and principles are published in Annals of Internal Medicine.
The working group used a consensus-based process to develop eight principles for using AI in a clinical guideline context. These principles are transparency, preplanning, additionality, credibility, ethics, accountability, compliance and evaluation. To ensure transparency, all AI tools should be clearly documented and made accessible to all users. The decision to use AI must be carefully planned before beginning the guideline development process and all members of the guideline development group must be made aware of the advantages, risks and limitations. The use of AI in the guideline enterprise should also bring about more benefits than risks, and the AI tools used should demonstrate sufficient quality and credibility. Using AI in guideline development must follow current ethical principles, and humans must oversee all AI outputs to ensure accountability. The use of AI must be compliant with legal and regulatory frameworks and continuous evaluation of the use of AI in the guideline enterprise based on these principles should be completed. Members of the working group suggest that anyone developing guidelines should refer to these principles and involve AI experts who can hold the AI tools and their outputs accountable during the guideline development process.
Media contacts: For an embargoed PDF, please contact Angela Collom at [email protected]. To speak with corresponding author Holger J. Schünemann, MD, MSc, PhD, please email [email protected].
