Do freelancers have to tell clients they use AI?
In many cases, yes, and where it isn't strictly required it's still the safer move. On a bonus episode of Being Freelance, host Steve Folland talks with lawyer Ingrid Fernandez about five ways AI use can quietly put a freelance business at legal risk. Her advice comes down to transparency: tell clients which tools touch their information, check your confidentiality clauses before pasting anything into an AI, and let your contract do the heavy lifting.
From the episode
5 Ways AI Could Be Tripping Up Your Freelance Business - Bonus - with lawyer Ingrid Fernandez
Aug 23, 2026 · 33 min
Client personal data and your privacy policy
The first trap is the most common. A freelancer drops a client's email thread, a customer list, or meeting notes into an AI tool to summarise or draft something. That's personal information being processed by a third party, and Ingrid says your privacy policy is supposed to cover exactly this.
Her position is that clients should be told about every tool that handles their personal information, and AI tools belong on that list alongside your email provider and your project management app. She's candid that nobody fully knows what happens inside large language models, which is why she also recommends turning on whatever controls the tool offers to stop your inputs being used for training. If you're using AI for note-taking on client calls, say so up front rather than leaving it to be discovered.
"We should be disclosing to people all the tools that touch their personal information."
Material clients send you might be AI-generated too
The second risk runs the other way. Clients are increasingly handing freelancers briefs, copy, images, and reference material that they generated with AI, sometimes without saying so. If you build on that material and it turns out to infringe someone's rights or can't be protected as IP, you can end up in the middle of the problem.
Ingrid suggests simply asking clients whether the source material was created with AI, and adding a clause to your contract that requires them to tell you. It's a small question that shifts responsibility back to where it belongs.
Using AI in your own service delivery
The third and most searched-for issue is disclosure of your own AI use. Ingrid's framing is practical rather than absolute: figure out where you're using AI and what you're using it for, then be transparent about it in relation to the services and products you deliver. A copywriter using AI to draft first passes is a different situation from a designer using it for a background texture, but in both cases the client should be able to find out.
Her argument is that transparency protects you. If a client later objects to AI being involved, the fact that you disclosed it in your terms or proposal is your defence. If you hid it, you have none.
"It's really about recognising where you're using it and what you're using it for."
Who owns what the client does with your work
The fourth risk is newer. Once you deliver work, clients may feed it into AI tools to generate variations, extensions, or entirely new derivative pieces. Ingrid says your contract needs to define what rights the client gets over the deliverable and whether creating AI derivatives is included.
This matters more for illustrators, photographers, writers, and designers whose style is the product. If you don't want your work used to train a model or to spin out endless variations, the contract is the only place that restriction has any force.
"Your contract is where it sits. If something goes wrong, we wanna have a mechanism."
Confidentiality clauses now name AI specifically
The fifth risk is the one that can get you fired. Ingrid says she's seeing a clear trend in client contracts: confidentiality clauses that explicitly prohibit putting the client's information into AI tools. If you signed one of those and then used an AI assistant to summarise their strategy doc, you've breached the contract, regardless of how harmless it felt.
Her advice is to go back and read the confidentiality sections of every active client agreement, and to treat client information with the same care you'd want for your own. She closes on a reassuring point: a well-drafted contract should make all of this simpler, not harder, because the rules are settled once and you stop second-guessing every task.
"We are really seeing now a lot more confidentiality clauses are being strict about AI specifically."
What to remember
- List AI tools in your privacy policy alongside every other tool that touches client personal data, and turn off training on your inputs where possible.
- Ask clients whether the material they send you was AI-generated, and put that obligation in your contract.
- Be transparent about where and how you use AI in delivering your services; disclosure is your protection.
- Define in your contract whether clients can create AI derivatives of your work.
- Re-read confidentiality clauses in existing agreements; many now ban AI use explicitly.
People also ask
Is it illegal to use AI on client work without telling them?
It depends on your contract and privacy obligations. If a confidentiality clause bans AI or their personal data is involved, undisclosed use can be a breach. Ingrid's advice is to disclose regardless.
How do I disclose AI use to clients?
Update your privacy policy to list AI tools, mention AI use in your proposal or terms, and tell clients proactively if you use AI for call notes or drafting.
Can a client use my work to train AI or make variations?
Only if your contract allows it. Ingrid recommends spelling out the client's rights over deliverables and whether AI-generated derivatives are included.
Based on the bonus episode "5 Ways AI Could Be Tripping Up Your Freelance Business" with lawyer Ingrid Fernandez, released August 23, 2026 on Being Freelance, hosted by Steve Folland.
