Be Good · Be Fair · Be Kind

AI & Gig-Work Fairness Policy

Be Good · Be Fair · Be Kind

1. Purpose

To ensure Able's AI systems and marketplace practices embody our core values: Be Good, Be Fair, Be Kind.

2. AI governance

TransparencyA plain-English summary of the factors our matching uses is published on this page. Users can request an explanation of any automated recommendation.
ContestabilityWorkers or buyers may trigger human review of a match or decision within 72 hours.
Bias monitoringWe monitor matching outcomes for bias on an ongoing basis and will publish headline fairness metrics as the platform scales.
Data minimisationOnly shift-relevant attributes — skills, availability and location — are passed into the model. The matching system never sees names, photos, age, gender, ethnicity or any demographic data.
Safety & securityModel inputs and outputs are logged and retained for 12 months for incident response; the platform undergoes regular security testing.

3. Fair-work standards

EarningsMinimum rate at or above the statutory National Minimum/Living Wage for the role; workers are paid 100% of agreed hours within 24 hours of shift completion.
AutonomyWorkers set their own rates, availability and travel distance; no exclusivity clauses.
TransparencyAll fees are displayed to both parties before booking; Able's 8% + VAT buyer fee is fixed and published.
RepresentationA community channel plus regular round-tables with the CEO.
Well-beingThe system flags sustained high workloads and prompts rest.
Dispute resolutionEscalation path: in-platform chat → Able support → independent mediation (CEDR model rules) within 28 days.

4. Recognition instead of ratings

No public star ratingsAble does not display 1-to-5 stars, streaks or public reputation scores for workers or buyers. Numerical scoring reinforces bias, causes stress and can entrench discrimination.
Recognition and development insteadAfter each shift both parties are invited to give structured feedback through short prompts ("What went well?" / "What could be improved?"). Feedback becomes recognition, training and development — visible only to the counter-party and Able's Trust & Safety function unless the author chooses to make it public.
AI-assisted summarisation, human-reviewedWhere many shifts occur, a natural-language summary helps users surface common themes; a human moderator checks tone and personal-data disclosure before release. Summaries can be contested via our community channel.
Use of feedback in matchingQualitative insights may be converted into a binary "trusted partner" flag (internal only) after manual verification. The flag can increase the weighting of future matches but never blocks access to shifts; any negative pattern triggers human review.
Right to challengeWorkers and buyers can request deletion or redaction of feedback they consider unfair or inaccurate, via the dispute path in §3.

5. Monitoring & reporting

We track key indicators — fill rate, cancellation rate, wage compliance and fairness metrics — and will publish them as the platform scales.

Last updated: 20 August 2026