AI Disclosure & Ethics

FleshWeave is an AI-assisted, human-directed studio, and we say so plainly. We believe AI should be disclosed, not disguised. This page explains, honestly, how we use it — so audiences, creators and partners always know what they are seeing and hearing.

AI-ASSISTED · HUMAN-DIRECTED

That tag is not marketing. It is a commitment: a person directs, finishes and signs off every piece of work that leaves the studio.

What AI does

• Accelerates drafting — arrangement ideas, rough mixes, variations and iteration that would otherwise take days.
• Assists production and post — clean-up, stem separation, mastering support, captioning and visual concepting.
• Helps maintain consistent, clearly-labelled educational creator personas.

What people do

• Set the brief, make the taste and editorial calls, and curate.
• Perform, write, and finish the work to release standard.
• Take authorship and accountability, and sign off before anything ships.

The machine is a fast collaborator. It is never the author. Human creativity is the point of FleshWeave; the AI is the loom, not the weaver.

AI-assisted work carries a visible provenance tag across the site, and we are building toward attaching durable content credentials so that disclosure travels with the file itself.

• On-site labels — a tasteful, machine-and-human readable “AI-assisted · human-directed” tag on relevant work.
• C2PA-style intent — we intend to adopt C2PA Content Credentials so provenance metadata (who, what tools, what was AI-assisted) is embedded and verifiable.
• Platform norms — where we publish to third-party platforms, we use their AI-content labelling controls.

People come first. Any persona, voice or likeness we work with is built on a foundation of explicit, informed consent.

• We do not create synthetic likenesses or voices of real people without their clear, documented permission.
• No unauthorised voice cloning. Ever. Cloning a voice requires the rightsholder’s consent and a written agreement.
• Our showcase personas are illustrative archetypes — moods, not real individuals — and are clearly marked as placeholder.
• Consent can be scoped and withdrawn; we honour reasonable withdrawal requests.

We aim to use tools and datasets that respect creators’ rights.

• We favour licensed, owned or permission-based material for any fine-tuning or reference.
• We respect copyright and will respond to good-faith rights concerns — email ai@fleshweave.com.
• We do not pass off AI-assisted work as wholly human, nor human work as fully autonomous AI.

We design our disclosure practices to align with emerging transparency standards, including:

• the EU AI Act transparency obligations for AI-generated and AI-manipulated content (Article 50), including clear labelling of synthetic media;
• UK guidance and the direction of travel on AI transparency and online-safety expectations;
• mainstream platform AI-labelling norms for synthetic and AI-assisted media.

This is a living commitment: as standards mature, our practices and this page will evolve with them.

• Impersonate real people or deceive audiences about authorship.
• Clone voices or likenesses without consent.
• Generate harmful, hateful or unlawful synthetic media.
• Hide AI involvement to make work look more “authentic”.

If you ever see something on the Platform that looks mislabelled or missing its provenance, tell us at ai@fleshweave.com and we will look into it.