>AI for Working Life

Review & Status

This dashboard presents the complete specification as a click-through prototype for stakeholder review. Content is drawn strictly from the programme brief; no figures, findings or instruments have been invented.

What the concept does well

Treats AI anxiety as a legitimate signal about real conditions (job security, surveillance, autonomy) — not a personal deficiency to be trained away.
Separates AI education from AI adoption, and names informed non-use as a valid outcome.
Builds in verification, privacy, human accountability and refusal skills from the start, not as add-ons.
Distinguishes individual skill-building from organisational conditions — and refuses to promise anxiety reduction or employability gains.
Applies implementation science (EPIS) and mixed-methods evaluation with cautious, explicit progression criteria.

Decisions required before pilot launch

Funding & protected time: no pilot can run ethically without paid or protected learning time confirmed per host organisation. Budget figures are structure-only (£TBC).
Jurisdiction: employment-rights, data-protection and scam-reporting signposting must be localised (see Referral & Escalation).
Instruments: all proposed measures are provisional, locally developed. A literature search for validated instruments is required before baseline.
Worker representation: union or employee-rep involvement must be agreed at exploration stage, not consulted after design.
Tool neutrality: demonstration tools will change mid-pilot; the curriculum must remain tool-agnostic at core.

How to use this dashboard

Navigate via the sidebar (tap the ☰ menu button on mobile). Curriculum modules open with tabbed detail. The Quality-Control Report checklist is interactive and saves locally. Anything marked [source verification required] must be verified before publication. Anything marked OPEN awaits a stakeholder decision.

Scope statement (binding): This programme is educational. It is not psychological treatment, employment advice, legal advice, financial advice, or a guarantee of job security. Facilitators signpost; they do not counsel, represent, or adjudicate.

Established findings Plausible / context-dependent Emerging Insufficient evidence Speculative — excluded from teaching

Deliverables Index A–W

Click any deliverable to jump to its section. All 23 requested deliverables are represented.

Programme Principles

Twelve binding design principles. Every curriculum and delivery decision must be traceable to at least one.

Ethics & Safety Statement

Non-negotiable commitments applying across all settings and pathways.

Scope limits (state at recruitment, at session 1, and in all materials): This programme is educational only. It is not psychological treatment, employment advice, legal advice, or a guarantee of job security. No participant outcome is promised.

Domain	Commitment
Informed consent	Plain-language consent covering purpose, data use, voluntariness, and the right to withdraw without consequence. Re-consent at evaluation timepoints.
Voluntary participation	No workplace mandate. Enrolment, attendance and all activities (including sharing) are voluntary. Non-participation must carry no documented penalty.
Privacy & data minimisation	Collect the minimum. Practice uses fictional, synthetic or public data only. No participant real data in demonstrations.
Confidential workplace information	Explicit session rule: never paste real confidential material into any AI tool during the programme — or afterwards, absent policy permission.
Cybersecurity	Teach account hygiene and scam awareness as part of safety, not as a barrier to entry.
Accessibility & adjustments	Reasonable adjustments offered proactively (see Accessibility & Equity section), budgeted, not on request only.
Vulnerable adults & safeguarding	Named safeguarding lead per host; facilitator escalation protocol; no unqualified counselling.
Job insecurity & distress	Content warnings before job-impact material; opt-out without explanation; referral routes ready before session 6.
Discrimination & surveillance	Acknowledged as legitimate concerns; route to union, HR policy channels, or data-protection complaint — never dismissed in-session.
Professional accountability	Consistent message: a named human remains accountable for decisions and outputs. AI never transfers accountability.
Limits of AI-generated advice	Health, legal and financial AI outputs are treated as unverified until confirmed by an authoritative human source.
Data retention & deletion	Published schedule; evaluation data pseudonymised; re-identification risk assessed, especially in small cohorts.
Equity of access	Device, connectivity, travel, childcare and language support provided — see Accessibility & Equity.

Target Population & Needs Assessment Deliverable B

Design for the widest realistic range of adults; differentiate within cohorts rather than screening people out.

Participant groups

Employed and self-employed
Unemployed and returning to work
Older workers
Workers in automation-affected occupations
Adults with disabilities
Migrants and multilingual learners
Adults with limited formal education
Adults with limited device, broadband, time or support access

Needs-assessment plan (Exploration phase)

Short survey per host organisation: digital confidence, AI exposure, top concerns, access barriers Provisional instrument
4–6 interviews or small groups per setting, including at least two people unlikely to volunteer for "AI training"
Organisational scan: existing AI policy, automation plans, learning-time norms, union/rep structures
Digital-access audit: devices, connectivity, quiet space, assistive technology
Concerns mapped to response level: individual / organisational / legal / wellbeing — the programme only owns the first

Inclusion/exclusion stance: exclusion criteria are minimised. Where a need is outside scope (e.g. acute distress, active employment dispute, literacy below course level), the response is referral plus an alternative offer, not removal from the programme.

Baseline digital-access requirement (minimum viable)

Access to a supported device during sessions (provided where lacking)
Basic point-and-click confidence — otherwise routed to a foundation digital-skills partner first, with re-entry guaranteed
No home broadband required: offline-capable materials and session-time connectivity provided

Objectives & Competency Framework Deliverable C

Learning outcomes across four competency strands. All outcomes are statements of intended learning, not promised results.

Strand 1 — Understanding AI

AI fundamentals and system types
Capabilities and limitations
Hallucination and output reliability
Bias and omission
Misinformation and synthetic media

Strand 2 — Safe & critical practice

Verification routines
Privacy, security, data classification
Prompt and task specification
Appropriate workplace use — and non-use
Human accountability

Strand 3 — Work & identity

Job and task analysis
Task change vs whole-job replacement
Professional identity
Workplace governance and employee voice
Role-specific safe-use or non-use plan

Strand 4 — Agency & life

Everyday-life applications and consumer AI
Perceived control (taught without overpromising)
Support-seeking
Personal AI learning plan

Calibration principle: confidence is assessed against judgement quality (knowing when to use, check, refuse, escalate) — never against enthusiasm for AI. Confidence without verification competence is treated as a risk indicator, not a success.

Pilot Model Deliverable D

A single design, delivered in at least two contrasting settings, before any scale decision.

Parameter	Specification	Status
Target population	Mixed-confidence adults from two contrasting host settings (e.g. one workplace, one community/adult-college)	Setting selection open
Recruitment channels	Employer/HR and union channels (workplace); college, library, employment services, community organisations (community)	Confirmed
Duration	8 weeks · 8 core sessions × 2 hours + 2 follow-up coaching sessions + drop-in clinic	Proposed
Delivery balance	Workplace cohorts: primarily face-to-face, 1–2 sessions online. Community cohorts: blended per access audit	Proposed
Group size	8–12 participants	Proposed
Facilitator ratio	2 facilitators per cohort (≈1:6) — one leads content, one supports access and wellbeing	Proposed
Pathways	Foundation · Workplace · Job-transition · Everyday-life · Advanced (see Pathways)	Confirmed
Follow-up & coaching	Coaching at +2 and +5 weeks post-programme; open drop-in to week 12; optional 3-month check-in	Proposed
Scheduling	Outside peak workload; repeated cohort times; recorded/recapped core content for missed sessions	Confirmed
Paid / protected time	Minimum condition for workplace delivery. Employer agreement in writing before enrolment opens	Per-host decision
Device & connectivity	Loan devices and data support available; no participant purchase required	Confirmed
Referral routes	Agreed in writing with partners before session 1 (see Referral & Escalation)	To confirm
Minimum safe-implementation conditions	Trained facilitators (2); safeguarding lead named; referral routes live; synthetic practice data prepared; venue accessibility checked; consent materials approved	Confirmed

Contingency plans

Scenario	Response
Low attendance	Below 5 in a session: merge cohorts or run condensed format; below 40% cohort enrolment uptake: pause, re-run needs assessment before relaunch
Technical failure	Every session has an offline, discussion-based equivalent; printed packs; no session is tool-dependent
Participant distress	Pause content; private check-in; referral per protocol; session can be restructured without the affected element
Policy conflict (e.g. employer restricts AI mid-pilot)	Curriculum adapts to focus on non-use, governance and voice modules; escalate to steering group; venue-neutral option

Modular Curriculum Deliverable E

Eight modules. Select a module, then use the tabs for full detail. Evidence ratings apply to the content basis, not to the effectiveness of this specific teaching — which is untested until the pilot.

Differentiated Learning Pathways Deliverable F

All pathways share Modules 1 and 8. Entry is by self-assessment plus a short conversation — never by test.

Pathway	For whom	Core modules	Emphasis
Foundation	Low digital confidence; limited prior exposure	1, 2, 7, 8 (+ bridging support)	Safety, basics, confidence without pressure
Workplace	Employed; AI appearing in current role	1, 2, 3, 4, 5, 6, 8	Policy, verification, task mapping, voice
Job-transition	Unemployed, returning, or in affected occupations	1, 2, 3, 6, 7, 8	Realistic task landscape, identity, referral to employment services
Everyday-life	Priorities outside work: scams, misinformation, services	1, 2, 3, 4, 7, 8	Consumer AI, synthetic media, official sources
Advanced	Confident users; team leads, reps, peer champions	3, 4, 5, 6, 8 + governance content	Critical evaluation depth, supporting others, governance literacy

Non-use pathway note: every pathway ends with a role-specific safe-use or non-use plan. A fully reasoned plan to not use AI, with verification of where that stance depends on conditions, is marked as a complete, successful outcome.

Facilitator Specification & Preparation Deliverable G

Required competencies

Adult-education practice with mixed-level, mixed-language groups
Working AI literacy — honest about their own limits and errors
Facilitating emotionally loaded topics without counselling
Privacy and data-protection basics
Neutrality on adoption decisions; comfort with silence and disagreement
Accessibility practice (adjustments, alternative formats)

Preparation plan (Preparation phase)

2-day training: curriculum, facilitation of distress, boundary-keeping (education ≠ advice ≠ advocacy)
Co-facilitation of one full cohort alongside an experienced lead before solo delivery
Monthly facilitator supervision/reflective practice session
Scenario bank rehearsed: "Am I going to lose my job?", "My employer is tracking us", "Is this scam real?", "I refuse to use this" — each with a facilitator response pattern and a referral boundary
Facilitator manual with fidelity indicators and adaptation log

Boundary rule: facilitators answer emotional and technical questions within education. Anything touching employment rights, mental health, legal position or discipline is signposted the same day, using the live referral routes — never handled by personal opinion.

Recruitment & Inclusion Plan Deliverable H

Barrier	Mitigation
Time poverty / shift work	Protected-time agreement; repeat cohorts at varied times; short recaps for missed sessions
Devices & connectivity	Loan devices; data stipends; offline materials
Language	Plain English throughout; key-term glossary translated to locally relevant languages; multilingual peer support where feasible
Disability & neurodiversity	Proactive adjustment offer at enrolment; quiet options; no forced group sharing; materials in alternative formats
Mistrust of "AI training"	Recruitment language leads with concerns and choice ("understand AI — use it or not"), not productivity promises; taster sessions
Self-selection toward the confident	Targeted outreach via unions, community orgs, employment services; monitor recruitment profile against the local population; revise channels if skewed
Care responsibilities & travel	Childcare support where needed; accessible venues; travel cost coverage

Workplace & Everyday-Life Application Framework Deliverable I

Task-impact mapping (taught tool)

List recurring tasks from a typical working week
For each: could AI change how it is done? who decides? what stays human?
Classify: unchanged / assisted / transformed / contested
Output feeds the role-specific safe-use or non-use plan — it is a personal analysis tool, not a workforce-planning instrument and never shared with employers without consent

Human–AI role boundaries

Named human accountability for every decision and output
Professional judgement stays with the practitioner; AI outputs are drafts or inputs
Explicit "do not delegate" list per role (e.g. disciplinary decisions, clinical/safety judgements, personal communications)

Organisational conditions — taught, not assumed away

Workload and job quality: AI may speed tasks without reducing hours — named as a governance question, not a personal one
Algorithmic management and surveillance: acknowledged as legitimate; routed to rep/union and data-protection channels
Organisational policy: participants learn to read, question and apply policy — including where policy is absent or unclear
Employee voice: consultation routes, union/rep involvement, collective as well as individual responses

Everyday-life coverage

Consumer AI assistants and personalisation
Misinformation and synthetic media: verification against official sources [local official sources to be confirmed]
Scams: cloned voices, fake messages, payment fraud — recognition and reporting routes
Health, financial and public-service information: AI output treated as unverified until confirmed
Escalation and support pathways for harm

Programme Logic Model Deliverable J

Read left to right. Outcomes marked "hoped-for" are explicitly not guaranteed and are tested, not assumed, in the evaluation.

Anti-overclaim rule: the logic model does not assert that AI literacy reduces anxiety, improves employment outcomes, or increases productivity. Those relationships are context-dependent and unknown for this programme until measured.

Implementation Roadmap (EPIS) Deliverable K

Mapped to the Exploration–Preparation–Implementation–Sustainment framework. Click each phase.

Implementation outcomes tracked: reach, acceptability, appropriateness, feasibility, fidelity, adaptation, cost, sustainability. Educational outcomes: knowledge, verification skill, privacy awareness. Behavioural: follow-up use/non-use decisions, support-seeking. Workplace: dialogue quality, policy clarity (organisational outcomes reported, not claimed as individual effects). Possible unintended outcomes monitored: overconfidence, blanket refusal without reflection, increased worry, misuse of programme as adoption mandate, equity gaps in reach.

Indicative Timeline Deliverable L

Twelve-month pilot cycle. All dates provisional pending funding confirmation. Click a phase.

Mixed-Methods Evaluation Framework Deliverable M

Timepoints: BaselineMidpointPost4–6 weeks3 months. All instruments below are provisional and locally developed unless a validated instrument is identified and cited pre-baseline [source verification required].

Domain	Type	Measure (provisional)	Timepoints

Qualitative strands

Participant interviews / focus groups at post + follow-up
Facilitator reflections after each session
Structured observation of 2 sessions per cohort
Adaptation logs (what changed, why, by whom)

Data management

Pseudonymised IDs; re-identification risk reviewed for small cohorts
Missing-data plan: report patterns; no imputation in pilot; interpret with caution
Subgroup analysis (descriptive only, powered for nothing): setting, pathway, baseline confidence, access level
Mixed-methods integration: joint displays mapping qualitative explanations onto quantitative patterns

These thresholds are proposed for stakeholder ratification — they are not empirically derived.

Completion ≥ 60% of enrolled participants across cohorts
Acceptability: majority of completers rate content and tone as appropriate
No serious adverse incidents attributable to the programme
Fidelity: ≥ 80% of core components delivered as specified (with adaptations documented)
Reach: participant profile broadly reflects the recruitment pool across confidence and access levels
Observable improvement on the verification-skill task rubric (descriptive)
No pattern of increased harm, worry, or pressured adoption in qualitative data

No-go signals: evidence of harm, equity failure, or organisational misuse of the programme as an adoption mandate halts scale-up regardless of other criteria.

Fidelity & Adaptation Checklist Deliverable N

Core components (not adaptable without steering-group approval)

Concerns acknowledged before benefits, every cohort
Education–adoption separation stated at recruitment and session 1
Informed non-use presented as legitimate
Verification practice in Modules 3 and 4
Synthetic/public data only for practice
Human accountability message in every pathway
Referral routes live from session 1

Adaptable periphery (documented in adaptation log)

Examples, cases and occupation-specific materials
Demonstration tools (version-logged)
Session order within modules; timing; venue
Language of glossary support; format of activities
Depth of governance content by audience

Adaptation documentation: every deviation recorded as: what / why / who approved / effect observed / retain-revert-review decision at pilot end.

Risk Register Deliverable O

Every risk has a named owner role and a contingency. Filter by category.

ID	Risk	L×I	Mitigation	Contingency	Owner

Accessibility & Equity Plan Deliverable P

Area	Provision
Digital accessibility standard	All materials target WCAG 2.2 AA; tested with screen reader and keyboard-only navigation before use
Reasonable adjustments	Offered proactively at enrolment and re-offered mid-programme; budget line reserved; no diagnosis or proof required for standard adjustments
Neurodiversity	Predictable session structure; agenda visibility; no cold-calling; quiet participation options (written, drawn, or silent reflection)
Language diversity	Plain English; translated glossaries; multilingual peer support; facilitator guidance on mixed-language facilitation
Age & education diversity	No assumed prior knowledge; varied formats (spoken, visual, hands-on); glossaries without condescension
Devices, connectivity, costs	Loan devices, data support, travel and childcare coverage — all pre-budgeted, not means-tested at point of need
Equity monitoring	Reach profile reviewed against local population at midpoint; corrective recruitment action if skewed

Referral & Escalation Plan Deliverable Q

Routes must be named, live and agreed in writing before session 1. Local entries below are placeholders pending confirmation. [local routes to be confirmed]

Concern raised	Route	Timing
Distress, anxiety affecting wellbeing	Named wellbeing/signposting contact per host; GP or local support service as appropriate	Same day
Employment rights, redundancy, discipline	Union rep / recognised union; ACAS or equivalent advisory body [verify jurisdiction]	Same day
Discrimination concerns	Union/rep route; equality advisory body; organisational grievance route if participant chooses	Within 48h
Surveillance / data misuse by employer	Organisational DPO; national data-protection authority complaint route [verify jurisdiction]	Within 48h
Scams or fraud exposure	National fraud/cyber reporting route; bank contact for financial loss [verify local route]	Immediate
Safeguarding concern (vulnerable adult)	Host safeguarding lead per agreed protocol	Immediate
Employability support needs	Employment services partner; careers guidance	On request
Technical access needs	Programme device/connectivity support	On request

Facilitator rule: signposting is the intervention. Facilitators do not accompany, represent, counsel or advise beyond education — even when asked.

Budget Structure Deliverable R

Structure only — all amounts £TBC pending funder and host confirmation. All figures open

Category	Components	Amount
Staffing	Programme lead (fractional); facilitator delivery + paid preparation time; accessibility support worker	£TBC
Participant support	Device loans; data stipends; travel; childcare; backfill for protected learning time where employers cannot cover	£TBC
Venue & technology	Accessible venues; session connectivity; demonstration devices	£TBC
Materials	Plain-language and translated materials; alternative formats (large print, audio); printing	£TBC
Evaluation	Evaluator time (independent preferred); participant thank-you vouchers; transcription	£TBC
Partner time	Referral partner capacity (union reps, wellbeing contacts) — agreed, not assumed free	£TBC
Contingency	10–15% for adaptations, rescheduling, additional adjustments	£TBC

Scale-Up Decision Framework Deliverable S

Four sequential gates. Failing any gate stops scale-up; it does not necessarily stop the programme.

Cautious by design: no gate can be passed on enthusiasm, anecdote, or vendor claims. Every gate requires documented evidence from the evaluation framework.

Stakeholder Questions Deliverable T

Assumptions Requiring Confirmation Deliverable U

Assumption	Owner	Confirmation method
Funding covers participant support, not only delivery	Sponsor / funder	Budget sign-off
Employers will provide paid/protected learning time	Programme lead + employer	Written agreement per host
Union / worker-rep engagement is achievable early	Programme lead	Exploration-phase meeting record
Host AI policies exist or are developing (enables realistic teaching)	Host liaison	Organisational scan
Referral partners have capacity to receive signposts	Partnerships lead	Capacity agreement
Validated instruments exist for key constructs (trust calibration, perceived control)	Evaluation lead	Literature search pre-baseline [required]
Jurisdiction-specific signposting routes can be named	Compliance liaison	Local verification
Demonstration tools will remain accessible without paid accounts	Curriculum lead	Tool check before each cohort

Evidence & Citation-Verification Note Deliverable V + W (part)

Verification status of this document: this specification was produced without live source verification. No citations, statistics or instruments have been fabricated. Every factual claim requiring support is marked [source verification required] and must be verified before external publication or funding submission.

Evidence tier	Definition	Treatment in this programme
Established	Well-supported general findings	Teach as reliable background (e.g. generative systems can produce confident errors)
Plausible / context-dependent	Reasonable basis; varies by setting and population	Teach with conditions stated (e.g. psychological safety supports learning)
Emerging	Early or limited evidence	Teach as "current best understanding, likely to change" (e.g. prompting technique research)
Insufficient	Little relevant evidence	Do not teach as fact; frame as open question
Speculative	Unsubstantiated — including deterministic, promotional and dystopian claims	Excluded; explicitly named as speculation where participants raise it

Proposed search strategy for verification [to execute]

Peer-reviewed: adult-learning and andragogy literature; occupational-health psychology; digital-inclusion research; implementation science (EPIS and related frameworks)
Systematic reviews: AI literacy interventions; digital-skills training outcomes; misinformation-susceptibility
Official guidance: national data-protection authority; equality/accessibility bodies; employment advisory bodies
Labour market: credible statistical and research bodies — prioritising task-based (not occupation-level) analyses
Exclusions: vendor whitepapers treated as claims, not evidence; preprints flagged as such
Each accepted citation recorded with: claim supported, population, setting, and whether it concerns training or organisational change — these are never conflated

Standing prohibitions honoured throughout: no correlation-as-causation; no claim that literacy automatically reduces anxiety; no guarantee of employability; no inevitable-jobs narrative in either direction; confidence ≠ competence; trust ≠ safe use.

Quality-Control Report Deliverable W

All 22 checks from the brief, with status. Click items to mark your own verification — progress saves in this browser.

Verification progress0%

Draft status is shown alongside each item; your personal ticks are independent.