AI Systems Transparency & Compliance

All services offered on Lovvai (including Photo Studio, Photo Book, Morphify, Remix, Caption Studio, and Image Editor) utilize advanced generative artificial intelligence models invoked via API. Users are explicitly notified that every image, portrait, or caption produced constitutes a synthetic output generated or manipulated by artificial systems, not a live photograph. In open-prompt services such as Photo Studio, Lovvai operates strictly as a neutral technological tool acting on user instructions, without offering third-party catalogs and without a general obligation of preventive monitoring over text prompts (Art. 8 EU Reg. 2022/2065).

Pursuant to Article 50(2) of the EU AI Act and European Commission Guidelines, every media file generated and downloaded from the platform embeds machine-readable technical metadata (IPTC Photo Metadata, dedicated EXIF tags, and C2PA Content Credentials standards) certifying the synthetic origin of the image. Pursuant to Art. 50(4) for creative works, disruptive visible watermarks are not applied to the visual surface upon private download.

The user holds exclusive editorial control over submitted text prompts and generated images. Pursuant to Article 50(4) of the EU AI Act, if the user decides to publish, share on social media, or externally distribute an image generated on Lovvai that depicts real people or appears photorealistic (deepfake), the user assumes the legal role of public Deployer and is required by law to clearly and visibly disclose the artificial origin of the content. Using or distributing likenesses of third parties or public figures without consent for commercial, promotional, or reputationally harmful purposes remains strictly prohibited (Art. 10 Italian Civil Code and Artt. 96-97 Law 633/1941), as does the criminal offense of transferring, publishing, or distributing deceptive deepfakes causing unjust harm (Art. 612-quater Italian Penal Code, Law 132/2025).

Uploaded user selfies are processed solely to preserve the user's facial likeness during the requested generation, strictly on Explicit Consent (Art. 9.2.a GDPR). Under Recital 51 GDPR, graphic generation from a text prompt does not involve biometric identification or profiling of third parties. For B2B clients, DISINDEX S.R.L. acts as Data Processor (Art. 28 GDPR). Transient session images are deleted within 24 hours, and no user image or prompt is ever used to train public or third-party AI models (EDPB Opinion 28/2024).