Transparency
Last updated: July 2026
I advise on responsible AI. So you can expect me to set an example myself. On this page you can read exactly how I use AI in my work, what happens to your data, and when I deliberately leave AI alone.
Anthropic’s Claude is my main AI assistant: for research, writing, code and for building client solutions such as chatbots. I use OpenAI’s ChatGPT alongside it, mostly to compare and check results. I use Perplexity for research that needs current sources, and Mistral’s Le Chat as a European alternative, the same provider I recommend as a greener option in my CO₂ calculator. I transcribe conversations and voice notes with tools such as Whisper and tl;dv, and only when everyone involved knows about it. Automation tools such as Zapier link workflows together, with AI as one link in the chain, never an end in itself.
For every client project I say which AI tools are part of the solution, so nothing about it comes as a surprise later.
Confidential client data doesn’t go into public AI models: no client names, no financial details, no personal data belonging to your clients in a prompt. I use AI services on settings or subscription plans where input isn’t used to train models.
If we work together over a longer period, we put the use of AI in writing: a data processing agreement with an AI addendum setting out which tools, which data and which limits. And everything I make for you, with or without AI, is yours, including the prompts and instructions written specifically for your business.
With me, AI does the preparatory work and I make the decisions. Everything that reaches you or your clients has been checked and reviewed by me, and I verify any factual claim from AI against the original source before I use it. The advice is mine: AI helps me think, but my judgement, my experience and my responsibility stand behind every recommendation.
I leave AI out of sensitive conversations and personal situations: those deserve a person, not a model. I leave it out when a simpler solution is better, because not every problem needs AI, and sometimes a well-built spreadsheet or a clear process is the honest answer. And I leave it out when the privacy cost is too high to justify. In that case I’ll say so, and we’ll look for another way together.
Using AI costs energy and water. I keep my own use deliberate: I pick the lightest model that can handle the task, and I don’t reach for AI on work that goes perfectly well without it. Curious about the climate impact of your own AI use? That’s what I built a free AI CO₂ calculator for (in Dutch).
Transparency works both ways: ask me anything. Email info@kimberleyvanruiven.nl, or see the privacy policy for how this website itself handles data.