European Commission's (HTACG): General Principles for Using Artificial Intelligence in the Preparation of Joint Clinical Assessments Dossiers

Sharan Murugan
Jul 20
3 min read

Artificial Intelligence (AI) is increasingly being explored across evidence generation, analysis, and reporting, offering opportunities to improve the efficiency of processes supporting health technology assessment. However, when AI is used to prepare evidence for regulatory and assessment purposes, its potential benefits must be balanced against risks to completeness, methodological quality, scientific rigour, and reliability.

To address these considerations, the Member State Coordination Group on Health Technology Assessment (HTACG) adopted the General Principles on the Use of Artificial Intelligence in the Preparation of Dossiers for Joint Clinical Assessments on 15 July 2026. The document provides general principles for situations in which AI is used to prepare evidence synthesis for Joint Clinical Assessment (JCA) dossiers under Regulation (EU) 2021/2282 on Health Technology Assessment.

The objective is to ensure that the scientific quality of the steps leading to a JCA is maintained even when AI-assisted methods are introduced.

Scope of the Principles

The document provides general principles for the use of AI in preparing evidence synthesis included in JCA dossiers and includes all models, including generative models, developed through machine learning, regardless of the specific model architecture.

The principles recognize that AI may be used across different stages of evidence synthesis, while emphasizing that its use does not change the existing legal obligations or methodological requirements applicable to JCA dossiers.

JCA Dossier Requirements

The Health Technology Developer's (HTD's) dossier forms the basis of the Joint Clinical Assessment.

Under Regulation (EU) 2021/2282, the dossier must be complete regarding available studies and data that could inform the assessment. The evidence must also be analysed using appropriate methods to address all questions within the assessment scope.

When AI is used to support dossier preparation, the Health Technology Developer must continue to comply with these requirements and follow the methodological guidance established by the HTACG.

Accountability for the Use of AI

Health Technology Developers remain accountable for the content, methods, and findings of the evidence synthesis presented in the JCA dossier.

This responsibility includes decisions about whether and how AI is used, understanding its impact on evidence synthesis, and ensuring the validity of outputs generated or informed by AI.

HTDs are also responsible for ensuring that AI use complies with applicable legal requirements, including copyright and data protection obligations, particularly because the JCA dossier will be published. The use of AI systems should also be consistent with applicable requirements and guidance under Regulation (EU) 2024/1689, the AI Act.

Human Oversight

Human oversight should be maintained throughout all AI-assisted steps involved in preparing the JCA dossier.

No step in dossier preparation or analysis should be fully automated without a human ultimately responsible for its quality and accuracy. This principle reinforces that AI can assist the evidence synthesis process but does not replace the accountability and scientific judgment of the individuals responsible for the dossier.

The use of AI also does not change the applicability of existing HTACG methodological guidance. Where AI is used, it should remain possible to verify that the methods applied comply with the relevant requirements.

Transparency in AI Use

Transparency is a central principle when AI contributes to the preparation of a JCA dossier.

If AI is used, this should be clearly documented in the dossier. All AI-assisted steps should be specified, including activities such as:

Information retrieval and study screening
Data extraction
Risk of bias assessment
Analysis
Reporting

Any outputs or judgments that are AI-generated or AI-informed should also be transparently identified.

However, the use of AI solely to improve spelling and grammar does not need to be disclosed.

Reporting AI Use in the JCA Dossier

The dossier should provide specific information about any AI tools used during its preparation.

At a minimum, the information reported should include:

Name of the AI tool
Version and date of the tool
Developer of the AI tool
Specific purpose or purposes for which the tool was used

As a good practice, any adaptations made to commercially available AI tools should also be described.

In addition, all prompts used with AI tools are expected to be recorded and made available if requested during the Joint Clinical Assessment process.

By emphasizing accountability, human oversight, transparency, and detailed reporting, the principles support responsible AI adoption while protecting the scientific quality and reliability of evidence used in Joint Clinical Assessments.