AI Transparency
MorphMe Live produces AI-generated and AI-transformed video. This page explains, in plain terms, how that content is marked and disclosed — for our users, their viewers, regulators, and our AI provider Decart. We publish this so anyone can verify our approach in minutes.
1. What is AI-generated here
MorphMe Live transmits your webcam feed to Decart's Lucy 2.1 Realtime model (decart.ai), which transforms your appearance into a fictional character in real time. The output you stream or record is therefore synthetic, AI-transformed video. MorphMe Live does not generate any content itself — all transformation is performed by Decart's model, and all input comes from you.
2. How AI content is marked and disclosed
We treat AI disclosure as layered, not a single switch. The layers currently in place:
2.1 Provider provenance mark (Decart watermark)
Decart, as the AI model provider, applies a provenance watermark to realtime output. Under the EU AI Act, the provider of a generative system is responsible for marking synthetic output in a machine-readable, detectable way (Article 50(2)). We do not remove, obfuscate, or interfere with Decart's mark, and our Acceptable Use Policy §4.G and Terms §6.2.E forbid users from stripping it.
2.2 Viewer-facing disclosure (Stream Kit)
MorphMe Live ships an in-app Stream Kit (Settings → Stream) that provides ready-to-paste AI-disclosure language for Twitch titles, Twitch/Kick panels, and YouTube descriptions — disclosure baked into every template variant. This implements the deployer-side disclosure obligation for AI-transformed content (EU AI Act Article 50(4)). Our use case is evidently fictional and creative character transformation (anime, original, and archetype characters), the category Article 50(4) recognises for proportionate disclosure.
2.3 Consent attestation on reference images
The first time a user uploads a reference image, MorphMe Live presents an Upload Consent Modal requiring a three-point attestation: that the likeness is fictional or consented, that AI use will be disclosed to viewers, and acknowledgement of our AUP. A persistent reminder appears on every subsequent upload.
3. Our commitments (in progress)
We are strengthening machine-readable provenance in our own output, independent of the provider watermark:
- Embedded content credentials — we are working to embed machine-readable AI-provenance metadata (C2PA / content credentials) into recordings so that AI provenance travels with the file, not only on screen.
- Persistent in-output AI badge — a lightweight, always-present "AI · MorphMe Live" marker so disclosure survives in any downstream use of the content.
We will update this page as each ships. If you are a regulator, platform, or rights-holder and need detail on timing, contact us at hello@morphmelive.com.
4. What we prohibit
AI transformation is powerful and we draw hard lines. Impersonating a real person without verifiable consent, sexual content involving minors, non-consensual intimate imagery, fraud, election interference, and the full EU AI Act prohibited-use catalogue are banned. The complete list is in our Acceptable Use Policy, which mirrors Decart's AUP clause-by-clause (see AUP §8 for the mapping table).
5. Report a concern
If you see content made with MorphMe Live that violates these principles, report it at morphmelive.com/report or email hello@morphmelive.com. We acknowledge valid reports within 14 days.
6. Regulatory framing
We design for the EU AI Act (Article 50 transparency obligations, in force 2 August 2026), Nigeria's NDPA 2023, and the disclosure rules of the streaming platforms our users broadcast on. AI provenance and disclosure are not afterthoughts for us — they are how a synthetic-media product earns the right to operate.
This page complements our Acceptable Use Policy, Privacy Policy, and Terms of Service.
THE TOOLS HUB — Federal Republic of Nigeria — CAC Business Name Registration No. 9204311
AI-transparency enquiries: hello@morphmelive.com