AI Usage Policy
Desilenced Design
Version: 1.0 Last updated: 16 June 2026
Desilenced Design is an AI-native creative studio based in London, United Kingdom. We use artificial intelligence as part of our creative production workflow to develop cinematic advertising, brand films, music videos, AI avatars, UGC-style content, still imagery, campaign visuals, social assets, scripts, edits, motion concepts and related brand content for clients in the UK, the EU, the US and other markets.
This AI Usage Policy explains how we use AI tools, what clients can expect, what clients are responsible for, how rights and liability are allocated, and what types of work we will not create.
1. Scope and relationship to other documents
This policy should be read alongside our Privacy Policy and Terms of Use (where published), and alongside any proposal, project contract, usage licence or written agreement between Desilenced Design and the client.
If there is any conflict between this policy and a signed proposal or project contract, the signed proposal or project contract prevails. Where a matter is not addressed in a signed contract, this policy applies.
This policy sets out expectations, responsibilities and the basis on which we work. It is not a substitute for the project contract, which contains the binding commercial terms for each engagement.
2. Our position on AI
Desilenced Design uses AI as a creative production tool, not as a replacement for creative judgement.
Our work is directed, edited and finished by hand. AI may support image generation, video generation, storyboarding, concept development, animation, voice exploration, visual research, editing, captioning, upscaling, compositing, production planning or creative iteration. The final creative direction, judgement, selection, editing and delivery remain human-led.
We do not present AI as magic, and we do not rely on raw AI output as final professional work. Our value is in creative direction, concept development, campaign thinking, visual consistency, editing, production control and final delivery.
3. How we may use AI tools
Depending on the project, we may use AI tools and related software for:
concept development and campaign ideation
scripts and voiceover drafts
image and video generation
animation and motion prompts
AI avatar development and character consistency systems
product and brand visualisation
music video keyframes, lip-sync and performance development
UGC-style ad concepts and social content systems
background generation, visual research and moodboards
image enhancement or upscaling
video editing support, captioning and formatting
delivery assets for social media, websites and advertising
The exact workflow varies by project. AI tools, editing tools and production platforms may change over time as technology develops.
4. Human direction and final review
All work delivered by Desilenced Design is subject to creative review before delivery. We aim to check outputs for visual quality, brand alignment, obvious AI artefacts, identity consistency, misleading or unrealistic claims, inappropriate likeness issues, obvious intellectual property concerns, platform suitability, commercial usability, and final formatting and delivery requirements.
AI outputs can contain errors, distortions, artefacts, unexpected details or visual inconsistencies. We use reasonable skill and care to reduce these issues, but no AI-assisted workflow can guarantee perfect first-generation results.
5. Client responsibility for supplied materials
When you provide materials to Desilenced Design, you confirm that you have the right to provide and use them for the project. This includes, where relevant: brand assets, logos, product images, music, lyrics, scripts, voice recordings, photographs, video footage, personal likenesses, employee or founder images, influencer or creator references, customer testimonials, third-party creative references, location images, campaign claims, product claims, business information and competitor references.
You must not provide materials that infringe copyright, trade marks, privacy rights, publicity rights, confidentiality obligations, employment obligations, platform rules, advertising rules or any other third-party rights.
If you provide images, voices, names, identities or likenesses of real people, you confirm that you have obtained the necessary consent, release, licence or lawful authority for us to use those materials in the agreed project.
You are responsible for the accuracy and lawfulness of everything you supply, and for the claims and instructions you give us. We are entitled to rely on the materials and approvals you provide.
6. Real people, likenesses and voice
We do not knowingly create non-consensual deepfakes, impersonations or misleading synthetic representations of real people.
If a project involves a real person, including a founder, employee, artist, influencer, customer, actor, model or public figure, the client must confirm that the person has authorised the relevant use, unless the use is clearly lawful and agreed in writing. This applies to face references, body references, voice references, AI voice cloning, AI avatar creation, lip-sync generation, testimonial-style content, founder videos, employee-led content, influencer-style content, digital doubles, synthetic presenters and music artist identities.
You are responsible for holding any required consents, releases and publicity-rights permissions for the territories in which the content will be used. We may refuse or stop work where consent, ownership or usage rights are unclear.
7. AI avatars and synthetic presenters
Desilenced Design may create AI avatars, fictional presenters, digital models, AI influencers, character systems or synthetic campaign talent.
Unless agreed otherwise in writing, AI avatars created by Desilenced Design are intended to be fictional or brand-owned characters, not deceptive copies of real people. Where an avatar incidentally resembles a real person, that resemblance is unintended.
For AI avatar projects, we may create character identity sheets, angle sheets, expression sheets, wardrobe direction, visual consistency prompts, voice or script direction, UGC-style videos, social content systems, campaign stills, website visuals and ad-ready assets.
Clients must not use AI avatars in a way that falsely suggests a real person has endorsed, used, reviewed or recommended a product or service unless that person has genuinely done so and the necessary rights and permissions exist. Where an avatar or presenter could be mistaken for a real person, the client is responsible for any disclosure required under section 8.
8. Disclosure and transparency
We support clear disclosure of AI use where it is legally required, contractually required, platform-required, ethically appropriate or necessary to avoid misleading the audience.
We do not automatically label every AI-assisted asset as AI-generated, in the same way that ordinary post-production is not separately disclosed. However, disclosure is more likely to be necessary, and in some cases legally required, where:
an AI avatar, presenter or voice could be mistaken for a real person
synthetic likeness or voice is used
a testimonial-style asset is fictional or performed by an avatar
the content could mislead viewers about who participated
platform rules require AI disclosure
advertising rules require transparency
the content deals with health, finance, legal, political, medical or other sensitive areas
EU. From 2 August 2026, the EU AI Act introduces transparency obligations for AI-generated and manipulated content, including a duty on those who deploy content that constitutes a deepfake to disclose that it has been artificially generated or manipulated, and obligations relating to machine-readable marking of synthetic outputs. Where we deliver content for use in the EU that may fall within these rules, we will apply or support appropriate disclosure and provenance marking, and the client is responsible for ensuring the required disclosure is present when the content is published. Platform-applied labels do not, on their own, satisfy these obligations.
UK. UK advertising rules apply to content regardless of how it is produced. There is no blanket requirement to disclose AI use, but disclosure does not cure a fundamentally misleading message. A claim that misrepresents a product, result or experience is not made acceptable by adding a note that AI was used.
US and elsewhere. Disclosure, endorsement and synthetic-media rules vary by country and, in the US, by state and by platform. The client is responsible for reviewing whether published content requires disclosure in each channel, country and context where it is used.
Where disclosure is needed, it can be simple and brand-safe, for example: "Created using AI-assisted production", "AI-generated brand avatar", "Fictional AI presenter", "AI-assisted visual campaign" or "Synthetic character created for brand storytelling".
9. Testimonials, endorsements and advertising claims
We do not knowingly create fake testimonials, fake reviews, false endorsements, fabricated customer experiences, false before-and-after results, fake medical claims, fake financial claims or misleading performance claims.
If a project includes claims about results, outcomes, pricing, safety, performance, revenue, health, beauty, medical treatment, investment, legal services, software performance or any other material claim, the client is responsible for ensuring the claim is accurate, substantiated and legally compliant in every market where the content is published. This is especially important for beauty and aesthetics, health and wellness, medical or treatment-related services, financial products, legal services, SaaS products, security tools, education products, coaching or consulting, supplements or regulated products, and paid advertising campaigns.
We may ask for evidence, clarification or revised wording before producing or delivering content containing claims.
10. Intellectual property and ownership
Ownership of deliverables. On full payment of all fees due for a project, Desilenced Design assigns to the client the rights it holds in the final bespoke deliverables created specifically for that project, so the client may use them for the agreed purposes. Until full payment is received, all rights remain with Desilenced Design.
What is not assigned. The assignment does not include, and Desilenced Design retains all rights in: our pre-existing materials, tools, methods, workflows and know-how; any AI prompts, systems or templates we develop; and our own characters, avatars, personas and in-house creative properties (including our in-house AI artists and influencer avatars). Any use of these by the client requires a separate licence agreed in writing.
Limitations of protection. AI-assisted outputs may attract limited, uncertain or no copyright protection in some jurisdictions. For example, the US Copyright Office does not register works that lack sufficient human authorship, and the position in the UK and other countries continues to evolve. Desilenced Design assigns only such rights as it actually holds in a given output and does not warrant that any specific deliverable is protectable, registrable or capable of exclusive ownership. The terms of the AI tools and platforms used may also affect the rights available in an output.
No copying of third-party work. Clients must not ask us to copy another brand's campaign, another artist's exact style, protected characters, celebrity likenesses without rights, copyrighted stills or footage, unauthorised music or lyrics, competitor branding, trade marks or logos without permission, platform-restricted content or confidential materials. References may be used to communicate mood, tone, lighting, pacing, framing, genre or colour, but not to produce confusingly similar copies of third-party work.
Portfolio rights. Desilenced Design retains the rights described in section 16.
11. AI tools, third-party platforms and model training
Some AI tools and creative platforms operate under their own terms of service, usage rules, licence conditions, restrictions and privacy settings. By commissioning AI-assisted work, the client acknowledges that certain project materials may be processed through third-party creative tools where reasonably necessary to deliver the project.
Some third-party tools may use the inputs they receive to operate, improve or train their own models. We take reasonable steps to use tools and settings that limit the use of client inputs for model training where such options are available, but we do not control third-party platforms and cannot guarantee how they handle inputs.
If your project involves confidential, unreleased, regulated or commercially sensitive materials that must not be processed by third-party tools, or must not be used for any form of model training, you must tell us in writing before work begins so that we can assess whether the project is suitable for our workflow and route it accordingly.
We aim to use tools responsibly and professionally, but we cannot control future changes made by third-party platforms, including changes to their pricing, availability, output quality, licence terms, model behaviour, moderation systems or service rules.
12. Data protection and privacy
Where a client supplies personal data, such as images, voices, likenesses, names or testimonials of identifiable people, the client is generally the data controller for that data and Desilenced Design acts as a processor handling it on the client's instructions for the purpose of delivering the project. A separate data processing agreement may apply where required.
Some supplied materials may include special-category or biometric data (for example, facial images or voice recordings of identifiable individuals). The client confirms that it has a lawful basis and any necessary consents for us to process such data for the agreed project.
Delivering a project may involve transferring materials to, and processing by, third-party tool providers that may be located outside the UK and the EEA. The client authorises such processing where reasonably necessary to deliver the project and is responsible for ensuring that any required transfer safeguards are in place for its own published use.
The client is responsible for compliance with data protection and privacy laws applicable to its content and audience, including UK GDPR and the Privacy and Electronic Communications Regulations in the UK, the EU GDPR in the EU, and applicable US state privacy laws (such as the California Consumer Privacy Act and equivalent state laws), together with comparable laws in other markets.
How we handle enquiry and contact data submitted to us through this website is described in our Privacy Policy.
13. Confidential and sensitive projects
Clients should tell us in writing before the project begins if any materials are confidential, commercially sensitive, unreleased, under NDA, legally privileged, subject to regulatory rules, subject to employment restrictions, involving minors, involving vulnerable people, involving sensitive personal data, involving medical, legal or financial claims, or restricted from being uploaded to third-party tools.
We may decline, adapt or restrict the workflow for projects that carry heightened legal, ethical, confidentiality or reputational risk.
14. Prohibited uses
Desilenced Design will not knowingly create or support content that is intended to:
deceive people about a real person's participation
impersonate someone without permission
create non-consensual sexual or intimate content
harass, exploit, defame or threaten a person
create fake evidence or fraudulent documents
produce false testimonials, fake reviews or misleading endorsements
spread harmful misinformation
misrepresent health, medical, legal or financial outcomes
infringe third-party intellectual property
bypass platform policies or legal restrictions
target protected or vulnerable groups unfairly
promote illegal activity
create political persuasion content using deceptive synthetic media
damage someone's reputation through manipulated or false media
We reserve the right to refuse, pause, revise or terminate work that we consider unsafe, unlawful, unethical, misleading, reputationally risky or inconsistent with this policy.
15. Client review and approval
The client is responsible for reviewing final content before publication. Client approval confirms that, to the client's knowledge: the content is accurate; the claims are substantiated; the required permissions exist; the content is suitable for its intended use; the content complies with relevant advertising, platform, data protection and industry rules in the markets where it will be used; the content does not infringe third-party rights; and the correct disclosure has been applied where needed.
Once final assets are approved and delivered, the client is responsible for how they publish, target, distribute, advertise, monetise, edit or repurpose the content, unless otherwise agreed in writing.
16. Portfolio and case study use
Unless restricted by written agreement, Desilenced Design may display selected project outputs, screenshots, stills, clips, process breakdowns, anonymised project notes or case studies in its portfolio, proposals, website, social media or marketing materials.
If a project is confidential, unreleased, sensitive or restricted, the client must state this clearly in writing before work begins. We will not knowingly disclose confidential commercial information, private client materials or sensitive behind-the-scenes details where confidentiality has been agreed.
17. Limitations of AI-assisted production
AI-assisted production can be fast and cost-effective, but it has limitations. Possible limitations include artefacts, visual distortions, inconsistent hands, faces, bodies or objects, imperfect lip-sync, continuity issues, unexpected details, platform moderation blocks, tool downtime, model behaviour changes, output variation between generations, lower control than traditional live-action production in some areas, and additional editing required to reach a professional finish.
Desilenced Design uses its judgement, workflow and post-production process to reduce these issues, but clients should understand that AI-native production is not the same as live-action production, 3D production or traditional photography.
18. Compliance across jurisdictions
Desilenced Design provides creative production services. We do not provide legal, regulatory, medical, financial or advertising-compliance advice unless expressly agreed in writing.
The client is responsible for ensuring that its final published content complies with all laws, advertising codes, industry rules, platform terms and market-specific requirements that apply to its business, product, service, audience and territory. Depending on where the content is published, this may include, without limitation:
UK: the CAP and BCAP advertising codes, consumer protection law, UK GDPR and PECR.
EU: the EU AI Act (including the Article 50 transparency obligations applying from 2 August 2026), the EU GDPR, and consumer protection law.
US: Federal Trade Commission rules, including those on endorsements, testimonials and fake reviews, applicable state privacy laws, and state right-of-publicity laws.
Other markets: equivalent advertising, data protection, intellectual property, image-rights, disclosure and consumer-protection rules.
We recommend that clients seek professional legal or compliance advice before publishing content in regulated or high-risk sectors, or in unfamiliar markets.
19. Warranties, liability and indemnity
We provide our services with reasonable skill and care. We do not guarantee specific results, performance, reach, rankings, conversions or that any output will be free of artefacts or perfect on first generation.
To the fullest extent permitted by law, the total aggregate liability of Desilenced Design arising out of or in connection with a project, whether in contract, tort (including negligence), breach of statutory duty or otherwise, is limited to the total fees actually paid by the client for that project.
To the fullest extent permitted by law, Desilenced Design is not liable for indirect or consequential loss, or for loss of profit, revenue, business, goodwill, anticipated savings or data.
Nothing in this policy limits or excludes any liability that cannot be limited or excluded by law, including liability for death or personal injury caused by negligence, or for fraud or fraudulent misrepresentation.
The client agrees to indemnify and hold Desilenced Design harmless against any third-party claims, losses, damages, liabilities, costs and expenses arising from or connected with: the materials, instructions, claims or approvals the client supplied; the client's publication, targeting, distribution or use of the deliverables; the absence of any consent, release, licence or disclosure that was the client's responsibility; or the client's breach of this policy or of applicable law.
20. Governing law and jurisdiction
This policy and any dispute or claim arising out of or in connection with it are governed by the law of England and Wales. The courts of England and Wales have exclusive jurisdiction to settle any such dispute or claim.
21. Updates to this policy
We may update this AI Usage Policy from time to time as AI technology, platform rules, legal requirements and our production workflow develop. The latest version, with its version number and update date, will be published on our website.
22. Contact
For questions about this AI Usage Policy, contact: