AI Practice Lab

Act as an experienced UK substance-use educator. Rewrite the source text below in calm, non-judgemental UK English for [AUDIENCE]. Aim for [READING LEVEL OR TONE]. Explain necessary clinical terms immediately in plain language. Preserve all cautions and uncertainties. Do not add new factual claims. Structure the answer with a short introduction, clear headings and practical key points.

SOURCE TEXT:
[PASTE NON-CONFIDENTIAL TEXT]

Create a [MINUTES]-minute learning exercise for [AUDIENCE] on [TOPIC]. The learning objective is: [OBJECTIVE]. It will be delivered [ONLINE / IN PERSON] to [GROUP SIZE] people. Use UK English and a non-judgemental, trauma-aware approach. Include facilitator instructions, a fictional scenario, three discussion questions, an answer guide and a short debrief. Mark any claim that requires source verification. Do not present the scenario as clinical advice.

Help me plan a UK evidence briefing on: [FOCUSED QUESTION]. The audience is [AUDIENCE]. First, break the question into six research questions. Then propose a transparent evidence-search plan prioritising official guidance, systematic reviews and primary research. Create a briefing structure with sections for established evidence, uncertainty, practice implications and sources requiring verification. Do not invent citations or claim that you searched material you cannot access.

Using only the verified source material I provide, draft a one-page harm-reduction resource about [TOPIC] for [AUDIENCE]. Use calm, direct UK English. Include: what people need to know, practical risk-reduction steps, signs of an emergency and where to seek help. Preserve uncertainty. Do not add dosage advice or unsupported claims. End with a list of statements that a qualified human reviewer must check before publication.

VERIFIED SOURCE MATERIAL:
[PASTE NON-CONFIDENTIAL CONTENT]

Summarise the document below for [AUDIENCE] and [PURPOSE]. Use only information in the supplied text. Organise the answer under: key message, important findings, implications for drug and alcohol practice, limitations, actions to consider and questions left unanswered. Add page or section references where available. Clearly label anything that cannot be determined from the document.

DOCUMENT:
[PASTE OR ATTACH APPROVED NON-CONFIDENTIAL MATERIAL]

Create a clearly fictional case study for [PRACTITIONER GROUP] to explore [LEARNING OBJECTIVE]. Set it in a generic UK service. Include enough ambiguity for discussion without creating a single “correct” response. Avoid stereotypes and moralising language. Include strengths, social context and more than one possible interpretation. Then provide five reflective questions and facilitator notes covering safeguarding, professional boundaries and uncertainty.

Draft a concise UK commissioning briefing about [TOPIC OR DECISION] for [AUDIENCE]. Use only the verified evidence and non-identifiable local context I provide. Structure it under: decision required, context, what the evidence shows, uncertainty and evidence gaps, equality considerations, delivery options, risks, questions for providers and recommended next steps. Distinguish clearly between evidence, local intelligence and assumptions. Do not invent figures, costs, legal duties or recommendations. End with a verification checklist.

VERIFIED MATERIAL:
[PASTE APPROVED NON-CONFIDENTIAL CONTENT]

Create a first-draft structure for a UK drug and alcohol service specification covering [SERVICE OR POPULATION]. Use only the approved information I provide. Include: purpose, population and inequalities, scope, access and referral, service model, safeguarding, co-occurring needs, lived-experience involvement, workforce expectations, partnership working, outcomes, quality measures, data and reporting, mobilisation and review. Mark every missing local decision as [TO CONFIRM]. Do not invent targets, activity levels, budgets, legal duties or clinical requirements.

APPROVED INFORMATION:
[PASTE NON-CONFIDENTIAL MATERIAL]

Help me plan a UK drug and alcohol needs assessment for [PLACE / POPULATION / DECISION]. Create: key questions; an evidence map separating prevalence, treatment, harm, inequality, access, outcomes and lived experience; suggested authoritative data sources to verify; stakeholder groups to involve; gaps and biases to test; and a transparent limitations section. Do not invent local findings or imply that national data describe local need. Flag small-number, disclosure and representativeness risks.

Create a transparent options-appraisal framework for [DECISION]. The options are [OPTIONS]. Organise the appraisal around the agreed criteria [CRITERIA], plus equality, implementation, workforce, safeguarding, evidence strength, service-user impact, risk and uncertainty. For each option, provide questions to answer and evidence required. Do not score, rank or recommend an option unless I supply verified evidence and an agreed scoring method. Keep assumptions separate and label missing information.

Draft a two-page leadership briefing on [TOPIC] for [BOARD / GROUP]. Use only the approved material supplied. Structure it as: purpose, decision or assurance required, current position, evidence, key risks, equality and service-user considerations, options, recommendation from the named accountable lead, actions and questions for the meeting. Keep facts, interpretation and unresolved issues separate. Do not invent assurance, performance, costs or decisions.

APPROVED MATERIAL:
[PASTE NON-CONFIDENTIAL CONTENT]

Create an AI-readiness discussion guide for a UK drug and alcohol organisation considering [PROPOSED USES]. Cover purpose, workforce confidence, information governance, data protection, approved platforms, procurement, equality, accessibility, safeguarding, output verification, accountability, training, incident response and review. For each area, provide: questions to ask, evidence of readiness, warning signs and a proportionate next step. Do not give legal approval or assume any platform is suitable.

Help me structure a small quality-improvement cycle for [SPECIFIC SERVICE PROBLEM] in a UK drug and alcohol setting. Using the non-identifiable information supplied, define the problem, map the current process, identify questions rather than assumed causes, suggest a small change to test, propose process and outcome measures, include balancing measures, equality considerations, staff and service-user involvement, review timing and stop criteria. Do not diagnose the cause or invent baseline data.

Analyse the approved, de-identified feedback below for [PURPOSE]. Group comments into provisional themes, include both positive and critical views, identify contradictory or minority perspectives, distinguish frequency from importance, and provide short paraphrased examples rather than traceable quotations. State what cannot be concluded because of sample, method or missing context. Do not infer prevalence, motive, diagnosis or causation.

DE-IDENTIFIED FEEDBACK:
[PASTE APPROVED CONTENT]

Create a set of reflective supervision questions for [ROLE / TEAM] exploring [PRACTICE THEME]. Use a strengths-based, trauma-aware and non-judgemental approach. Include questions about relationship, power, uncertainty, bias, safeguarding, boundaries, evidence, emotional impact, team support and next learning. Do not diagnose, decide a real case or imply a single correct response. End with a reminder to use established supervision and safeguarding routes where required.

Rewrite the verified referral pathway below for [AUDIENCE]. Use plain UK English and preserve the exact eligibility, exclusions, contact routes and escalation instructions. Structure it as: who the service is for, when to refer, how to refer, what happens next, urgent situations, accessibility and where to check for updates. Flag any gap or contradiction rather than filling it in. Do not add service promises or change thresholds.

APPROVED PATHWAY:
[PASTE CURRENT NON-CONFIDENTIAL CONTENT]

Create a [DURATION] multi-agency workshop plan on [TOPIC] for [PARTICIPANTS]. The purpose is [PURPOSE] and the decisions or outputs required are [OUTPUTS]. Include preparation, welcome and psychological safety, a concise evidence input, lived-experience contribution with appropriate support, structured discussion, areas of disagreement, decision points, actions, owners, review date and an accessible follow-up record. Build in ways for quieter voices to contribute. Do not use real case material.

Review the proposal below and generate equality questions for accountable human consideration. Explore who may benefit, be excluded or experience unintended harm; barriers to access and retention; digital exclusion; disability and neurodiversity; language and literacy; race and culture; sex and gender; age; sexuality; poverty, housing and justice involvement; geography; and how lived experience will shape decisions. Separate evidence supplied from gaps to investigate. Do not predict impacts or speak for communities without evidence.

PROPOSAL:
[PASTE APPROVED NON-CONFIDENTIAL CONTENT]