Content methodology and AI assistance
DosePractice has been developed with substantial AI assistance. AI has helped draft code, question templates, explanations, tests, documentation and website content. DosePractice remains responsible for deciding what is published.
AI output is not treated as authoritative. Human error, drafting mistakes and technical defects can also occur. The expanded bank is automatically checked for arithmetic consistency, but automated verification does not establish educational quality, regulatory status, clinical suitability or freedom from every wording error. Layered controls are used to reduce—not eliminate—the risk of error.
How the question bank is produced
- Question structures are limited to defined educational arithmetic domains.
- Values are generated deterministically so each question has a known answer and reproducible working.
- Automated tests verify identifiers, answer calculations, topic counts and privacy safeguards.
- Questions use fictional values and deliberately exclude medicine choice, contraindications, maximum doses, routes, patient factors and treatment recommendations.
- Before commercial publication or search indexing, the owner must complete substantive human review of the substance, wording, sources, boundaries and representative samples from every domain. External review may be added later if it is genuinely obtained and its scope is recorded.
Authorship and editorial responsibility
AI is not presented as the author or professional reviewer. A commercially promoted or indexed version must identify the responsible legal operator and record the owner approval status for each content set, the date checked, the sources used and the next review trigger or date. It must not imply an independent review team where none exists. Until owner review is completed, pages remain development drafts and search indexing stays disabled.
Why this disclosure is available
AI assistance does not remove the publisher's responsibility for accuracy, originality, transparency and useful educational content. Nursing-related educational material warrants clear authorship, source and review information. This page is linked from the Trust, FAQ, Terms and Content review pages rather than repeated beside every question.
What will not be done
- AI-generated material will not be published solely to create hundreds of thin search pages.
- AI will not be described as a nurse, doctor, regulator or independent reviewer.
- The disclosure will remain accessible through the Trust, FAQ, Terms and Content review pages.
- Automated checks will not be presented as legal, MHRA, NMC or university approval.
Controls used to limit error
- Formulae and worked steps are displayed so learners and reviewers can inspect the arithmetic.
- Deterministic question generation and automated tests check answers, identifiers, topic counts and regressions.
- Sources, versions, reviewer decisions and review dates must be recorded before commercial publication or search indexing.
- Substantive owner review must consider wording, ambiguity, units, rounding and educational boundaries; any future external review must be described only within its actual scope.
- Users can report a problem without creating an account, and material corrections must be logged, regression-tested and deployed through the rollback process.
No combination of human review or automated testing can guarantee that every error has been removed. Users must independently verify formulae, units and results using an appropriate trusted source.
Trust and transparency · FAQ · Content-review process · Privacy by design · Report a problem