Good AI Practice Principles 1–4: Foundation

The first four guiding principles establish the ethical, regulatory, and operational foundation for AI used to generate evidence in drug development. They apply whether AI supports toxicity prediction, clinical trial design, signal detection, or manufacturing analytics.

Principle 1: Human-centric by design

The development and use of AI technologies align with ethical and human-centric values.

For clinicians and researchers, this means AI should serve patient welfare and scientific integrity — not merely efficiency targets. Human oversight, accountability, and respect for patient rights remain central even when models automate substantial analytical work.

Hong Kong context

When reviewing AI tools for trial recruitment, dose optimisation, or adverse event coding, ask whether the design prioritises patient benefit and ethical conduct — consistent with local research ethics and Hospital Authority governance expectations.

Principle 2: Risk-based approach

The development and use of AI technologies follow a risk-based approach with proportionate validation, risk mitigation, and oversight based on the context of use and determined model risk.

Higher-risk applications — such as AI influencing primary efficacy endpoints or safety decisions — warrant more rigorous validation and oversight than lower-risk administrative uses.

Practical questions

• What is the model risk if the output is wrong?
• Is validation proportionate to that risk?
• Who provides oversight at each life cycle stage?

Principle 3: Adherence to standards

AI technologies adhere to relevant legal, ethical, technical, scientific, cybersecurity, and regulatory standards, including Good Practices (GxP) — such as GCP, GLP, and GMP as applicable.

For hospital pharmacists and quality professionals, this reinforces that AI is not exempt from existing quality frameworks. Data integrity, audit trails, and validated systems remain essential.

Principle 4: Clear context of use

AI technologies have a well-defined context of use — the role and scope for why it is being used.

Context of use clarifies:

• What decision or evidence the AI supports
• What it must not be used for
• How outputs fit into regulatory submissions or clinical workflows

Without a clear context of use, teams may over-extend a model beyond validated conditions — a common source of error in research and pharmacovigilance settings.

Clinical takeaway

Before adopting an AI tool, document its intended role explicitly. A model validated for signal prioritisation should not be repurposed for causal attribution without new validation.

Source: FDA & EMA — Guiding Principles of Good AI Practice in Drug Development (January 2026)