
AI IN MY WORKFLOW
Blog post
on AI
AI POLICY - ANTONELL MEDIA
Antonell Media is the trade name of Kajsa Antonell, sole proprietor of translation and copywriting services in Sweden. This policy explains how I use artificial intelligence in my client work.
HOW I USE AI
I use AI selectively and only when it serves a clear purpose: admin, structure, research, consistency checks and readability. It supports my process, it does not drive it. Whether I am working from a blank page or refining your AI-generated draft, the writing, translation and transcreation decisions remain entirely mine.
AI support does not change my responsibility for the work, and it does not automatically mean a lower price. Pricing is based on the assignment, scope, expertise, responsibility and value of the final deliverable, not on the tools used along the way.
TOOLS
I distinguish between standard support tools, such as spellcheckers, grammar suggestions, CAT tools, translation memories and termbases, and generative AI, meaning tools that generate or transform text based on prompts.
I may use generative AI for brainstorming, structure, background research or readability checks, and I may use private, controlled workspaces to organise or check material. I do not use client material as future reference material unless we have explicitly agreed to it.
AI translation and generative AI features in CAT tools are not part of my default workflow. I only use them when explicitly permitted by the client and project instructions.
RESPONSIBILITY AND QUALITY
I am personally responsible for all work I deliver. I review, edit and fact-check AI-assisted output. I do not treat AI as a reliable source. The final text is always written, edited, adapted or approved by me.
DATA PROTECTION
I treat all client material as confidential by default and handle personal data in line with GDPR and my privacy policy.
I do not enter sensitive material into public or unprotected AI tools, including personal data, unreleased campaigns, product names, strategy documents, legal material or anything under NDA.
When using AI tools, I use paid or private settings where input is not used for model training, whenever available. I also minimise, anonymise or generalise information where possible.
I do not authorise client material to be used to train any external AI model. I also do not use client material to train or expand my own tools, resources or reference systems unless we have explicitly agreed to it.
Antonell Media client material is never shared, transferred or repurposed for my art project Fulordsarkivet or any other separate project.
CLIENT REQUIREMENTS COME FIRST
Your requirements take precedence over this policy. If you have an internal AI policy, tool restrictions or specific requirements, I will follow them. This applies whether you want a stricter no-AI approach, a no-generative-AI workflow, or a project where you provide AI-generated material for me to edit, adapt or transcreate.
Unless agreed otherwise, I may use AI for low-risk admin tasks involving non-confidential material, meaning material that is not client-owned, unpublished, sensitive, internal or project-specific.
ONGOING REVIEW
I actively follow developments in AI tools, standards and professional practice, and update this policy as needed.
This Policy was last updated on APRIL 28, 2026.
For a more personal note on this policy and how I think about AI in my work, read the blog post.