AI features

Last updated: 7 September 2026

AI and automated processing

The App uses AI in two narrow, clearly bounded ways, and in no others. One runs entirely on your device: a species suggestion from your catch photo. The other is an automated image-safety check on photos that are public. Both are described below. The only image that ever leaves your device for an AI check is a photo that is public; nothing else you photograph, log or write is sent to an AI model.

• No chatbot and no AI assistant.
• Reports and verification requests are reviewed by a person, not by AI.
• No AI profiling of you, and no automated decision that produces a legal or similarly significant effect on you.

On-device species suggestion

When you log a catch with a photo, the App can suggest a species for it. This runs with a small model entirely on your device: the photo is not uploaded, sent to us, or sent to anyone else for it, and no internet connection is needed. The suggestion is only a starting point, and you always choose the final species. Nothing is decided about you from it, and the model ships pre-trained: it does not learn from your photos or catches. The suggestion is a Premium feature; without Premium the model never runs, and you pick the species yourself as always.

The suggestion can be wrong. It is an automated guess, not an identification you can rely on. Never use it to decide that a fish may legally be kept: species rules, protected statuses, and size limits are your responsibility to check against official sources (see the Terms, section 10).

Automated image safety check

When you make a catch public, its photo can be automatically screened for clearly objectionable content, such as explicit or graphically violent imagery, before it becomes visible to other people. We use Google Cloud Vision SafeSearch, an automated image-classification service, for this single check. It runs only on public photos: never on photos you keep private, and never on your text, measurements or location. Google states that images sent to this service are processed in memory and not stored.

If a photo is flagged it is withheld from public view while it is looked at; your catch still posts, without that photo, and you keep your own private copy. If it happens, we tell you, with the reasons and how to challenge it, and a person reviews the photo (email support@fishandbeer.app with the subject "Appeal"). This is an automated visibility measure to keep the shared feed safe, not a decision about your account. Like any automated check, it can miss things and it can flag harmless photos; that is why a person reviews flagged photos and why you can always report content you see. Legal basis: our legitimate interest in a safe service (GDPR Art. 6(1)(f)); the check is a voluntary own-initiative measure under Digital Services Act Art. 7, and Art. 17 governs how we tell you about any restriction.

Training on your content

Your photos and catches are not used to train AI models: not by us, and not by anyone else. We do not share them with a model provider for training, and we do not keep a training copy of them.

If that changes

We may add further AI features later, for example a more powerful species check that runs on our servers rather than on your device. If we ever send a photo of yours off your device to an AI model, this page is updated first, and we ask for your explicit permission before the first photo is sent. You would be able to decline and keep logging catches exactly as you do today.

Illustrations

The fish artwork in the species collection is stylised illustration, digitally processed for the App. It depicts the species generically. It is not a photograph of anyone's catch.

Questions

Anything unclear about this page: support@fishandbeer.app. Questions about how your data is protected can also go to legal@fishandbeer.app.