Applicability · Good AI Practice

Regulations/Good AI Practice/Applicability

/ Applicability logic

Applicability begins with scope, not assumption.

Good AI Practice

Applicability 01

Drug development AI

Discovery, nonclinical, clinical, manufacturing, PV, and regulatory decision-support may be relevant.

Applicability 02

Medicinal product lifecycle

EMA/HMA framing is not the same as FDA device GMLP, though themes overlap.

Applicability 03

Internal support AI

Applicability depends on whether outputs influence scientific, safety, quality, or regulatory decisions.

Applicability 04

Not a single binding code

Good AI Practice is an editorial workspace for regulator principles, not one harmonised regulation.

/ Dates and gates

Timing changes what can be responsibly claimed.

status gates

Gates

Track these before publishing a checklist or readiness claim.

2021 GMLP device principles comparator · 2024 EMA AI reflection paper · 2026-01 FDA/EMA Good AI Practice principles · model context-of-use decision · validation and monitoring plan

/ Use cases

Regulated use cases need a decision record.

life sciences lens

Use case 01

Context of use

Use this as a trigger for Good AI Practice applicability review, not as an automatic compliance conclusion. AI evidence starts with a precise problem, intended use, and decision context.

Use case 02

Data relevance and quality

Use this as a trigger for Good AI Practice applicability review, not as an automatic compliance conclusion. Data provenance, representativeness, completeness, and bias need review.

Use case 03

Model development

Use this as a trigger for Good AI Practice applicability review, not as an automatic compliance conclusion. Development choices should be documented and linked to intended use.

Use case 04

Validation and performance

Use this as a trigger for Good AI Practice applicability review, not as an automatic compliance conclusion. Performance evidence must fit context, population, endpoint, and risk.

Use case 05

Human accountability

Use this as a trigger for Good AI Practice applicability review, not as an automatic compliance conclusion. AI systems should not erase responsibility for decisions.

Use case 06

Transparency

Use this as a trigger for Good AI Practice applicability review, not as an automatic compliance conclusion. Users and reviewers need understandable information about limits and use.

/ Source anchors

Current public sources for Good AI Practice.

official first

These links are the public source anchors for this workspace. Interpretation, checklists, and future assets should point back here before being reused outside iFeed.

Official FDA page