AI Safety &
Ethics Policy

ICONISH is building the future of ecommerce content through AI-generated models and digital twins. We believe that the power of this technology comes with a profound responsibility. Our guiding principles: People First — AI should augment human creativity and opportunity, not replace human dignity or agency. Consent is Non-Negotiable — No one's likeness is ever used without explicit, informed, and documented consent. Transparency Over Obscurity — We are open about what our AI does, how it works, and what data it uses. Security by Design — Data protection is built into every layer of our technology. Accountability at Every Level — Humans are responsible for the outputs of our AI systems.

Our AI technology is used to generate original AI model personas (ICONS and KIN), create consent-verified digital replicas (TWINS), build bespoke AI talent (CUSTOM), and produce product-on-model imagery and campaign visuals. Our AI is not used for surveillance, behavioral profiling, political content, or any purpose beyond commercial content creation.

We do not train on client data. Materials you provide are used solely to fulfill your specific project — never to train AI models, never shared for model development, never used for other clients. We do not scrape social media or public databases to collect facial images. We do not purchase bulk facial image datasets from data brokers. We do not use facial recognition technology to identify or track individuals.

Our training data is licensed, consented, documented, and auditable. Our AI models are never trained on: scraped internet content; client-provided materials; images of minors; content from data brokers of uncertain provenance; or copyrighted material without proper license.

For every Source Individual whose likeness is used: • Written Consent — explicit written consent before collecting any biometric data. • Informed Consent — clear explanation of how their likeness will be used, by whom, for what duration. • Scope-Limited Consent — specific permitted uses, territories, and duration. • Compensated Consent — Source Individuals are fairly compensated. • Revocable Consent — withdrawal at any time; biometric data deleted within 30 days. The prohibition on creating Digital Twins of minors (under 18) is absolute with no exceptions.

We deliberately curate training data representing a wide range of ethnicities, body types, ages, and physical characteristics. We conduct regular bias audits and maintain human review checkpoints on every delivery. We do not produce content relying on harmful stereotypes or culturally insensitive imagery.

Every stage of our production pipeline includes human oversight. No content goes from AI generation to client delivery without human review and approval. Human creative directors review all briefs, model selections, and output specifications before AI generation begins.

We clearly communicate to all clients that the content we produce is AI-generated. We support clients in appropriately disclosing AI-generated imagery per applicable advertising standards. We maintain zero tolerance for deepfake creation, non-consensual intimate imagery, political disinformation, or deceptive content.

Our AI models are not trained on copyrighted material without proper license. Client materials are never used beyond the scope of the specific project, never used to train AI models, and securely deleted after project completion.

All data is encrypted in transit (TLS 1.3) and at rest (AES-256). AI compute infrastructure is hosted on SOC 2-compliant cloud environments. Regular penetration testing and comprehensive audit logging are standard practice. Biometric data is destroyed within 30 days of consent withdrawal or within 3 years of last interaction, whichever occurs first.

All technology partners with access to client data must execute data processing agreements prohibiting unauthorized use of that data, including using our inputs or outputs to train their own models.

We comply with state biometric privacy laws (BIPA, Texas CUBI), comprehensive state data privacy laws (CCPA/CPRA, VCDPA, CPA, CTDPA), state right-of-publicity statutes, and FTC guidelines on advertising disclosures and AI-generated content.

We prioritize cloud infrastructure providers that use renewable energy. We believe AI-generated content — by reducing the need for physical photoshoots, air travel, and wardrobe waste — offers an inherently more sustainable alternative to traditional content production.

If we identify any incident involving misuse of our AI technology or unauthorized data access, we will immediately investigate, notify affected parties, comply with all breach notification laws, and implement corrective measures. To report concerns, contact safety@iconish.ai.

ICONISH actively engages with industry bodies around AI safety and ethics. We are dedicated to regularly reviewing this policy, conducting periodic internal audits, and providing ongoing training to our team on responsible AI practices.

ICONISH Inc. AI Safety & Ethics: safety@iconish.ai Website: www.iconish.ai