How to Ethically Disclose Generative AI Assets in Client Deliverables (Without Losing the Contract)

The Silent Creep of Generative Assets

It starts small. You are working on a website mockup for a client, and you need a high-quality hero image of a generic modern office. Instead of spending hours hunting through stock photo sites or paying hundreds of pounds for a license, you open /platforms/midjourney, type a clever prompt, and generate a flawless image in forty seconds. It looks incredible.

But a quiet voice in your head asks: Do I tell the client?

Many freelancers and agencies are currently trapped in a cycle of silent AI usage. They fear that if they admit to using generative tools, the client will think they are lazy, demand a massive discount, or accuse them of plagiarism. Yet, staying silent is a high-stakes gamble. Intellectual property laws around AI-generated content are muddy at best, and if a client discovers you quietly slipped synthetic assets into their brand identity, the fallout can ruin your reputation.

Honesty doesn't have to cost you the gig. In fact, handled correctly, transparent AI disclosure can position you as a cutting-edge partner who knows how to deliver massive value efficiently.

1. Why Silence is a Legal and Ethical Risk

Before we look at how to disclose, we must understand why you have to. It isn't just about being a good person; it is about protecting your client’s business.

Currently, copyright offices in several major jurisdictions (including the US and UK) have ruled that purely AI-generated works cannot be copyrighted. If you deliver a brand logo or a key marketing asset that is heavily reliant on raw generation from Midjourney or Stable Diffusion, your client might not actually own the exclusive rights to it. If a competitor copies it, your client has zero legal recourse.

Furthermore, many enterprise clients have strict, non-negotiable clauses in their contracts regarding intellectual property and third-party tools. If you use generative AI without their knowledge, you might be in direct breach of contract. If they run your code through an automated audit and find traces of unlicensed synthetic code, or run your copy through an LLM detector (flawed as they are), the trust is broken instantly.

To understand the specific licensing and copyright implications of the tools you use, always keep an eye on official documentation and community support portals, such as the Midjourney Support guidelines, to ensure you know what you are legally passing on to your client.

2. Define Your Personal "Integration Scale"

To disclose AI use intelligently, you need to understand your own workflow. Not all AI usage is created equal. We can categorise AI integration into a simple three-tier scale:

Level 1: Editorial & Research Assistant (No Disclosure Needed)* Using LLMs to brainstorm ideas, proofread your own copy, find bugs in your hand-written code, or structure an outline. This is equivalent to using a advanced spellchecker or search engine. The final output is entirely yours; the AI was just a tool to help you get there. Level 2: Collaborative Production (Disclosure Recommended)* Using AI to generate substantial components of the work which you then heavily edit, refine, and merge with your own work. Examples include using /platforms/figma-weave to generate wireframe components that you then customise, or using a code assistant to write specific helper functions. Level 3: Pure Generative Assets (Disclosure Mandatory)* Delivering raw or lightly touched-up assets generated directly by an AI. This includes hero illustrations, full marketing copy drafts, synthetic voiceovers, or bulk-generated SEO landing pages.

By categorising your work this way, you can establish clear boundaries. You don't need to tell a client you used an LLM to proofread an email, but you absolutely must tell them if the key illustration on their homepage came out of a latent space model.

3. How to Frame Disclosure as a Superpower

The secret to disclosing AI use without losing your value is to change the narrative. If you frame AI as a way to "avoid doing the work," the client will undervalue you. If you frame AI as an accelerator that allows you to spend more time on high-level strategy and bespoke refinement, they will respect your efficiency.

Compare these two approaches:

> The Apologetic Approach: "Just so you know, I couldn't find a good stock image, so I had to use Midjourney to generate the hero graphic. I hope that’s okay."

> The Expert Approach: "To give you the most bespoke visual identity possible within your budget, I used custom-prompted AI models to generate a series of unique, tailored base illustrations for the website. I then manually edited, colour-corrected, and upscaled these in Photoshop to align perfectly with your brand guidelines. This approach saved us £2,000 in custom photography fees and allowed us to focus our budget on deep user-experience testing."

The second approach makes it clear that the AI is merely a raw material. The value you bring is your taste, your curation, your prompt mastery, and your post-processing skills. You are not charging them for the button press; you are charging them for knowing which button to press, and how to fix the output when it inevitably looks a bit weird.

4. Put It in Writing

Don't wait until the end of a project to drop the AI bomb. Make disclosure a standard part of your onboarding and invoicing process.

Include a simple "AI Usage Statement" in your project proposals. It can be as straightforward as:

> "We utilise generative AI tools (such as Midjourney and GPT-4) as part of our creative and technical workflow to accelerate research, ideation, and asset generation. Any synthetic assets used in final deliverables will be explicitly flagged, and we ensure all tools used comply with commercial licensing standards. Our human team remains solely responsible for the strategic direction, quality assurance, and bespoke refinement of all work delivered."

This simple paragraph sets clear expectations. It shows you are professional, modern, and legally literate. Clients who value their own intellectual property will respect this transparency, and it opens the door to a much healthier, collaborative relationship where nobody has to hide their browser tabs.