KaraokeCrowd is built and reviewed by people. We use AI in a few narrow places. This page explains where, what data is involved, and what we never use AI for. It follows the transparency rules of the EU AI Act (Regulation (EU) 2024/1689).
Where we use AI
The contribution chat assistant
When you add or correct a listing, you can use a chat assistant instead of the regular forms. The assistant is an AI language model, not a human. It turns your description (or an uploaded flyer photo) into a structured suggestion.
The assistant only drafts. Nothing is saved to the directory until you review the suggestion and confirm it yourself, and every confirmed suggestion then goes through the same review as a form submission.
Automated image checks
When someone uploads a photo, an automated service (AWS Rekognition, on EU servers) checks it for unsafe content. Anything the check cannot classify with confidence goes to a human moderator. The automated check never rejects an image on its own, and it does not identify people. It looks at content, not identity.
Finding karaoke leads (internal)
To find karaoke nights we might be missing, an AI model sorts public social media posts into likely karaoke venues, events, and hosts. Its output is only a lead. A person verifies every lead against our evidence rules before anything can become a listing.
We also sometimes use supervised AI agents to import listings from public venue lists. Their submissions go through the same review queue as everyone else's; nothing they file is published without review.
What we don't use AI for today
- Publishing decisions. AI never decides what appears in the directory. Every AI-drafted contribution is either confirmed by the person submitting it or reviewed by us before publication, and follows the same review rules as any other contribution. The rules that auto-publish some contributions from trusted members are hand-written code, not AI.
- The content you read. Venue pages, event details, help pages, and blog posts are written by people. We don't publish AI-generated text, images, or reviews.
- Profiling you. We don't use AI to profile you, score you, rank you, or predict what you might do.
- Personalizing what you see. We don't use AI to personalize what you see. Search and browse follow plain, predictable rules like distance and date.
- Decisions about your account. No decision with legal or similarly significant effect on you (like blocking an account) is made by automated means alone (Art. 22 GDPR).
- Emotion recognition or biometrics. No emotion recognition, no face recognition, no biometric identification or categorization. When the assistant reads a flyer, it is instructed to extract event facts and never to identify people in photos.
- Ad targeting or personalized pricing. We don't use AI to target advertising or to set personalized prices.
- Training AI models. We don't train AI models on your contributions, chats, or images.
If we one day add AI-generated content to the site — for example translated Memories or short atmosphere summaries — it will be clearly labeled as AI-generated, and this page will describe it before it launches.
What data reaches AI providers
The chat assistant sends the text you type, the contents of any files you attach, and a flyer image if you upload one to an external language-model provider so it can draft your suggestion. Please don't share sensitive personal information about yourself or others in the chat. Chat conversations expire 48 hours after your last activity and are permanently deleted shortly after; flyer images are not stored after reading. Which providers we use and on what legal basis is listed in our privacy policy.
The automated image check sends uploaded photos to AWS Rekognition on EU servers. Details are in the privacy policy as well.
Your rights and questions
The chat assistant is optional: every contribution can currently also be made through the regular forms.
If you have questions about our AI use, contact us through the details in our legal notice. Your data-protection rights (access, deletion, objection) are described in the privacy policy. In Germany, the market surveillance authority for AI systems is the Bundesnetzagentur; you can also raise concerns there.
We update this page whenever our use of AI changes.
