AI at Dateagent — Transparency Notice

Effective date: July 27, 2026

Dateagent uses AI in exactly one place, and this page explains how it works. There is no AI matchmaking, no AI ranking of people, no AI-generated messages sent on your behalf, and no chatbot.

Content screening: what it is

Dateagent never publishes your content on its own. There is no feed, no search, and no directory — your profile and date ideas become visible only through invitations you choose to share (links, QR codes, Wallet passes) or your datemenu page, which you can hide in Settings. Who you're dating, who you're planning dates with, and how those dates went stays private to your account, full stop.

The people you do share with can view invitation pages without accounts. So before shareable content becomes visible to them, Dateagent checks it against our Community Guidelines using an AI model. Screened content is exactly:

  • Photos: your profile photo (which can appear on invitation pages you share and on Wallet passes) and any contact photos you upload.
  • Public text: your profile's name, username, bio lines, and social handles; and each date idea's title, venue, address, hours, wardrobe, cost, and description.

How it works

  1. When you upload one of those photos or save one of those text items, Dateagent sends it to Anthropic, an AI provider, whose Claude model checks it against our published rules and returns an allow-or-block result with a short reason.
  2. If a photo is blocked, we never store it. If text is blocked, the save is rejected and the app shows you the reason so you can edit and try again.
  3. What is never sent: your private notes, date history, date feedback, contact names and phone numbers, and Guests' RSVPs. (Guest RSVP text is checked by a simple word filter on our own servers — it is not sent to any AI provider.)
  4. No facial recognition. The screening classifies what a photo depicts; it does not identify who is in it, and no facial-recognition template or biometric identifier is created or stored.
  5. Approved content is stored by Dateagent as ordinary app content. Anthropic's processing is transient, under its commercial API terms.

Your data is not used to train AI

Content sent to Anthropic through its commercial API is used to produce the screening result — not to train Anthropic's models, under Anthropic's commercial API terms. Dateagent also does not use your data to train any models of its own.

Its limits

Automated screening can be wrong in both directions — it may block something harmless or miss something it shouldn't. That's why reporting exists — in the app and at [email protected] — and why blocks aren't final: if something of yours was blocked and you believe it shouldn't have been, email [email protected] and a human will review it.

Your choices

  • Screening applies only to content your Guests could see, plus the photos you upload — what you choose to share is what gets screened.
  • Screening is a safety requirement of operating the Service and cannot be turned off. Your private content is out of its scope entirely.
  • Disagree with a block? Email [email protected] for human review.

Automated decisions

The only automated effect of this screening is on the individual photo or text you submitted — it becomes visible to your Guests or it doesn't. Dateagent does not use AI or automated processing to make decisions about you that have legal or similarly significant effects: no automated account suspensions, no scoring, no profiling for eligibility of any kind. Account enforcement and content-report handling are human-reviewed.

Questions: [email protected] · See also: Privacy Policy §5 · Terms of Service §13 · Community Guidelines