Skip to content

AI Ethics Policy

Where Apperah uses AI, how AI-generated content is labelled, and your choices.

Effective:
2026-06-29
Last updated:
2026-09-01
Version:
v1.3

Overview

Apperah uses AI for a limited set of functions: structuring publicly posted event information, drafting short descriptions of public places, automatically screening event listings and Open Invite photos users submit, and assistive features you actively use — trip itineraries and checklists, schedule-photo scanning, on-request translation, and the in-app feedback conversation. AI does not make decisions with legal or similarly significant effects about you; you can edit AI output about you; and our agreements with AI vendors prohibit the use of your personal data to train their general models.

This Policy explains where we use AI, where we do not, how AI-generated content is labelled, and the choices you have.

1. Where we use AI

  • Structuring public information — we turn publicly-posted event and venue information into clear, structured listings, and help search understand meaning. This uses public information only — not your personal data. Listings are screened automatically, and lower-confidence extractions are held back for review before they appear.
  • Checking that events are genuinely public — AI helps classify whether a publicly-posted event is open to everyone or restricted to a closed group (members, employees, students, private bookings). Restricted events are excluded from the public directory. This classification looks only at the public event text — never at people — and a fixed keyword check takes precedence over the AI’s judgement.
  • Keeping the platform safe — when you submit a public event listing, AI automatically reviews its text and images before it appears in the directory, and photos added to Open Invites are reviewed the same way. This review is part of publishing and cannot be switched off. We are additionally building automated child-sexual-abuse-material (CSAM) detection against authoritative registries. See Children’s & Youth Safety.
  • Helping you when you ask — some features send data to an AI provider only when you use them: generating a trip itinerary or packing checklist for a group journey, scanning a photo of a schedule you choose to turn into calendar entries, translating user-written content when you tap translate, and the feedback conversation in Messages, where an AI assistant may reply before our team follows up. Each runs on what you actively submit, and nothing else.

2. Where we do not use AI

  • No significant automated decisions about you. We do not make decisions producing legal or similarly significant effects about you by automated means alone (GDPR Art. 22).
  • Account actions are made by people. Suspending, restricting, or removing an account is a human decision, not an AI one.
  • Account-level moderation is human-reviewed. AI may help surface content at scale, and our automated safety review can decline a public event listing (you are told why and can contact support to contest it) — but any decision about your account is reviewed by a person.
  • Friend suggestions are rules-based — they derive from signals such as mutual friends and shared groups, not from an AI model.
  • No AI profiling of under-18s, and no ad-targeting profiling — we do not run behavioural profiling to target advertising. Promoted placements from businesses are selected by area and event category — never from a profile of you.
  • Recommendations are rules-based scoring, not an AI model — your home feed is ranked by signals such as distance, freshness, and your own in-app activity (see Privacy Policy §10), and you can switch personalised recommendations off in Settings > Data & personalization > AI Recommendations.
  • Analytics are not AI profiling. Our optional usage analytics (opt-in, pseudonymous — see the Privacy Policy) are used in aggregate to understand feature usage; they do not feed an AI model about you, and we do not build behavioural segments.

3. How we label AI-generated content

  • Content generated by AI — such as AI-drafted place descriptions — carries a clear “AI-generated” label in the app, together with a machine-readable marker (EU AI Act Art. 50, effective 2026-08-02).

  • Our Terms carry the matching Art. 50 disclosure, and we will publish a short “model card” for each AI feature on our website.

  • The label applies wherever AI-generated content appears, including inside groups and communities (for example an AI-drafted place description shown in a community’s recommended places).

4. Keeping you in control (human oversight)

  • AI does not make final decisions with a legal or similarly significant effect on you.
  • You can correct AI output — if an AI-drafted place or event detail is wrong, use Suggest an edit on the place or event and we review the correction.
  • Where AI flags something, a person reviews it before any action that affects an account.

5. Protecting your data when we use AI

  • Data minimisation — personal identifiers are minimised in content before it is processed by an AI provider, where feasible.
  • The search index is built from public venue text only — no personal data.
  • No training on your data — we require our AI vendors, under contract, not to use your personal data to train their general models.
  • We send personal data to an AI vendor only where there is a data-processing agreement in place and a documented lawful basis.

6. Prohibited practices

Apperah does not deploy the practices prohibited by the EU AI Act, including:

  • Subliminal or manipulative techniques, or exploitation of vulnerabilities, to distort your behaviour.
  • Social scoring.
  • Real-time remote biometric identification.
  • Untargeted scraping of facial images to build face-recognition databases.
  • Emotion recognition.
  • Biometric categorisation to infer sensitive attributes about you.
  • Predictive-policing systems.

7. The AI providers we use

We use third-party AI providers, grouped by task category:

  • Text understanding — extracting public event information, drafting place descriptions, importing venue menus from their public pages, and the automated safety review of event listings and Open Invite photos users submit.
  • Assistive features you invoke — trip itinerary and checklist generation, schedule-photo scanning, on-request translation of user-written content, and replies in the in-app feedback conversation.
  • Search — building a search index from public venue text so that search understands meaning (public data only).

We require a data-processing agreement with every AI provider that prohibits them from training their general models on your data, and we only send personal data to a provider once that agreement is in place. We name our providers on our Service Providers page.

8. Your choices

  • Turn AI-powered recommendations off in Settings > Data & personalization > AI Recommendations (see Terms §11). The automated safety review of public event submissions and Open Invite photos is part of keeping the platform safe and cannot be disabled.
  • The assistive features are yours to use or not — itinerary generation, schedule scanning, translation, and the feedback conversation each run only when you invoke them; not using them sends nothing.
  • Object to AI processing of your personal data — email privacy@apperah.com — and exercise your data-protection rights (including the right to object and the Art. 22 protections). See our Privacy Policy.

9. Our compliance commitments

  • EU AI Act Art. 50 (transparency) — effective 2026-08-02.
  • EU AI Act Art. 4 (AI literacy) — we provide appropriate AI-literacy training to staff who work on AI features.
  • We review each AI feature regularly and refresh our risk assessments when models change.

10. Contact

Questions or concerns about our use of AI: privacy@apperah.com.

11. Cross-references