How to Respond When a Client Says 'No AI on Our Work'
A client’s new contract arrives with a clause: “Supplier shall not use artificial intelligence tools in the performance of the Services.” Or a project manager writes: “Just so you know, we don’t want any AI used on our stuff.”
For freelancers, agencies and consultants, these restrictions are increasingly common. They are also usually broader than the client intends. Read literally, “no AI” could ban spell check, autocomplete, transcription and translation tools. A good response clarifies what the client is worried about, agrees something workable, and gets it in writing.
Understand What The Client Is Worried About
No-AI clauses usually come from one or more of these concerns:
- Confidentiality: their data being sent to a third-party AI provider.
- Quality: generic or inaccurate work.
- Ownership: uncertainty about who owns AI-generated material. See AI copyright in 2026.
- Their own clients’ requirements, passed down the chain.
- Regulation in their sector.
Each concern has a different fix. A confidentiality worry is solved by keeping their data off cloud services; a quality worry is solved by your process; an ownership worry may need a contract term.
Ask Before You Agree
Before:
Hi, re the no AI thing, does that mean I can’t use any AI at all? Like even Grammarly? Seems a bit extreme tbh
After:
Thank you for flagging the AI requirement. I want to make sure I follow it in the way you intend. Is the main concern that your information could be shared with third-party AI services, or that deliverables should be written without AI-generated content, or both? For context, I use a spelling and grammar checker, and for confidential material I use a writing tool that runs entirely on my own computer and sends nothing externally. I am happy to adjust my process to fit your policy.
The second version asks a real question, discloses your current tools honestly, and offers to adapt, without suggesting the client is being unreasonable.
Propose Workable Wording
If the clause is too broad, suggest wording that targets the real concern. For example:
Supplier shall not input Client Confidential Information into any AI system that transmits data to a third party, and shall not deliver content generated substantially by AI without Client’s prior written consent.
That protects the client’s data and deliverables while leaving room for tools like spell check and local processing. Have anything contractual reviewed by a lawyer; this is a starting point for discussion, not legal advice.
Professional bodies have also issued guidance on AI and client confidentiality. Lawyers, for example, have the American Bar Association’s Formal Opinion 512 (July 2024), and the Minnesota Lawyers Professional Responsibility Board’s article on ethics guidance for generative AI summarises the duties around confidentiality and informed consent. It is a useful reference for anyone working under professional obligations.
Keep A Record
Once agreed, confirm in writing what is and is not allowed, and keep it with the contract. If the client later asks how the work was done, you can show exactly what you agreed. See how to document your AI workflow for an auditor.
If The Answer Really Is “No AI At All”
Respect it. Turn off AI features in the tools you use for that client: AI writing suggestions in your email client, document summaries, meeting transcription. Keep a separate setup for that client’s work if needed. It is their data and their call.
Where A Local Tool Fits
Many clients’ real concern is confidentiality. A writing tool that runs entirely on your own computer, with no network calls during a rewrite, answers that concern directly, and is often acceptable when explained. But disclose it and ask; do not assume.
A Wrivio Context for client policy replies could say:
Rewrite this as a professional reply to a client about their policy. Respectful, cooperative and precise. Ask any clarifying questions clearly. Keep every tool, fact and commitment exactly as written. Do not promise anything that is not in the original and do not characterise the client’s policy.
Press Ctrl+Shift+Space, paste your draft, and make sure the result is honest about what you use.
Common Questions
Does a no-AI clause include spell checkers and grammar tools?
It depends on the wording. Many clauses are drafted broadly without thinking about everyday tools, so ask the client to clarify and confirm the answer in writing.
Can I use a local AI tool if the client bans AI?
Only if the client agrees. A local tool solves the data-sharing concern, but a ban may also be about how work is created. Explain your setup and ask.
Should I charge more for no-AI work?
If the restriction genuinely makes the work slower, it is reasonable to reflect that in your estimate. Explain it as a scope difference rather than a penalty.
What if I already used AI on their work before the clause?
Be honest. Tell the client what was used, how, and whether any of their data was shared. See how to tell a client you used AI on their work.
Download Wrivio for Windows to offer clients a private, on-device option for confidential work when they agree to it.
Read Next
What GPAI Enforcement Means for the Tools You Use
EU enforcement of general-purpose AI rules began in August 2026. You are not a model maker, but the rules still shape the tools you rely on. Here is how.
California's No Robo Bosses Act: What SB 947 Actually Requires
SB 947 would bar California employers from firing workers on an algorithm's word alone. What it covers, what it does not, and the deadline that matters now.
New AI Disclosure Laws in 2026: A Simple Way to Sort Them
California, New York, and the EU each added AI disclosure duties in 2026. Here is a three-question test to tell which ones actually apply to your work.
How to Run an End-of-Year Documentation Cleanup in One Afternoon
Stale wikis and outdated how-tos cost time all year. A one-afternoon routine to archive, update and fix your team's documentation before January.
This article is filed underPrivacy & Compliance, which has 96 articles.