AI Transparency Policy
Telloa can synthesize visuals and a consented presenter. The creator remains the editor and publisher of record.
1. Where AI is used
- Speech transcription, caption timing, and language processing.
- Research assistance, script drafting, storyboard planning, and quality checks.
- Image and video generation, visual replacement, and scene assembly.
- Optional rights-attested visual presenter generation and separately consented voice cloning.
- Automated rendering and delivery checks.
2. Human control
Your narration remains the default timing and performance authority. You can review and edit scripts, scenes, captions, and generated assets before publishing. Telloa does not publish a project to a social platform for you in the current product.
You are responsible for factual review, rights clearance, professional advice, and the final decision to publish.
3. Synthetic presenters and voices
Telloa gates visual presenter creation behind a versioned inline attestation that the uploader is the person shown or has that person’s explicit permission. This is a permission record, not formal identity verification. Voice cloning remains behind a separate explicit consent scope, and creator-recorded narration remains the default.
Do not use an enrolled presenter or voice outside the permission given by that person. Delete the enrollment to stop new Telloa generations tied to it.
4. When to disclose AI involvement
Clearly disclose artificial creation or manipulation when an output appreciably resembles a real person, place, object, or event and could reasonably be mistaken for authentic footage or audio. Keep the disclosure easy to notice and understand without unnecessarily disrupting an obviously creative or fictional work.
Also follow the rules of each publishing platform and any sector-specific duties that apply to advertising, endorsements, elections, finance, health, education, evidence, or journalism.
- Example: “AI-generated presenter; narration and editorial control by [creator].”
- Example: “Includes AI-generated visuals.”
- Example: “Synthetic reconstruction; not original footage.”
5. Limits
- Models can hallucinate facts, citations, code, visuals, and speaker behavior.
- Generation may reproduce stereotypes or produce unexpected material.
- Detection and provenance metadata can be removed by editing or platforms and should not replace a visible disclosure where one is needed.
- A generated output may resemble existing content without being an intentional copy.
6. Questions or reports
Contact hello@telloa.app about deceptive output or synthetic-media concerns.