This AI Disclosure explains how Lessimp Inc ("Lessimp," "we," "us") uses artificial intelligence across our mobile application, website, and related services (the "Service"). We built Lessimp on a Glass Box principle: you deserve to know why our AI made a decision, not just what it decided.
This disclosure is a companion to our Privacy Policy. The Privacy Policy describes data flows in general; this page focuses specifically on AI-driven processing, observability, and your rights around algorithmic decisions.
1. Agentic Transparency (Fiddler Integration)
Sovereign Principle: You deserve to know why our AI made a decision, not just what it decided. We operate a Glass Box, not a black box.
Lessimp uses Autonomous AI Agents to power matches, content recommendations, and safety moderation. To ensure these agents act fairly and without bias, we use Fiddler AI for Agentic Observability:
- Step-Level Tracing: Every decision made by our AI agents—whether suggesting a match or flagging content—is logged at the step level. This means we record not just the outcome, but the reasoning chain that led to it.
- The Glass Box Model: Unlike "black box" AI systems, our Glass Box approach means you can ask "Why am I seeing this?" for any AI-driven recommendation. Fiddler's observability layer provides the explainability infrastructure that makes this possible.
- Bias Detection: Fiddler continuously monitors our AI agents for drift, bias, and anomalous behavior. If a model begins producing unfair outcomes across demographic groups, the system alerts our engineering team and the model is retrained or corrected.
- What Fiddler Sees: Fiddler processes model telemetry—the internal reasoning of our AI agents. It does not receive your personal profile data, messages, photos, or financial information. Observability data is about how our agents behave, not about who you are.
2. Transparency Report
We publish the results of our AI fairness monitoring in our quarterly Transparency Report. This report includes aggregate statistics on AI decision accuracy, bias audit results, enforcement actions taken, appeal outcomes, and data request volumes. The report is available at lessimp.com and is updated every calendar quarter.
3. AI Sales Chat
When you engage with our chat widget on our public website (a floating chat bubble in the lower-right corner of every page), you may be interacting with our AI-powered sales assistant. We want you to know exactly how this works:
- Recorded and transcribed: The full text of your conversation, the timestamps of each message, the IP-derived general location, and the browser timezone you sent are stored in Firestore under
visitor_chats/<sessionId>. We retain active conversations until a Lessimp team member marks them resolved; once resolved, the transcript is permanently deleted within 90 days by an automated scheduled cleanup job (cleanupVisitorChats).
- Processed by Google Gemini: Your messages are sent to Google’s Gemini API for response generation. Google’s data handling is governed by the Google Cloud Terms of Service. Lessimp does not train or fine-tune its own AI models on these transcripts.
- You consent before sending: A checkbox in the chat widget requires you to acknowledge this recording before your first message is sent. You can close the widget at any time without sending a message — nothing is recorded if no message is sent.
- Human handoff: If the AI cannot answer your question or you ask for a human, we will ask for your email and a Lessimp team member at support@lessimp.com will reply by email.
- Right to delete: Email privacy@lessimp.com with your chat session ID (shown in the widget footer) and we will delete the transcript within 30 days.
4. AI Content Moderation
We use AI services to scan user-uploaded photos, videos, and text for prohibited content. Specifically:
- Image and Video: Photos and videos you upload are scanned by Google Cloud Vision API and Video Intelligence API for nudity, violence, hate symbols, and other policy-violating content.
- Text: Messages, profile content, and review text are analyzed for harassment, scam patterns, and policy violations. The AI flags suspected violations for human review — final enforcement decisions are made by a Lessimp team member, not the AI.
- AI-Generated Content Disclosure: When you use AI tools (e.g., Veo video reels) to create content on Lessimp, metadata about the generation is stored. Content shared publicly may include an AI disclosure label so other users know it was AI-generated.
- Reels and Marketing Content: Some of Lessimp's own marketing content (including hero videos and short-form reels on TikTok and YouTube Shorts) is AI-generated. These posts include an
#AIGenerated tag where appropriate.
5. Your Rights Around AI Decisions
If you are subject to an AI-driven decision on Lessimp (a match recommendation, a content moderation action, a safety flag), you have the following rights:
- Right to Explanation: You can request the reasoning trace for any AI-driven decision that affects you. Email privacy@lessimp.com with the date and context of the decision.
- Right to Human Review: If you disagree with an automated decision, you may appeal it to a human reviewer. The appeal queue is handled by Lessimp team members within 7 business days.
- Right to Object: You may opt out of certain AI-driven personalization through the Privacy & Data settings in the app. Opting out may reduce match quality or content recommendation accuracy but does not affect your access to the Service.
- Right to Restrict: Under GDPR Article 22, EU residents have the right not to be subject to a decision based solely on automated processing. Lessimp's match recommendations are assistive, not determinative — you always make the final choice to message or meet someone.
For questions about how Lessimp uses AI, or to exercise any of the rights described above:
This AI Disclosure is effective as of May 25, 2026, and is governed by Lessimp's Sovereign Global Standard. It is a companion document to the Privacy Policy and supersedes prior AI disclosure statements embedded in earlier versions of the Privacy Policy.
