Statement of position on AI use

This statement explains when, whether and how Crystal Clarity Copywriting Ltd (I) will – and will not – use artificial intelligence (AI) in the course of providing editing and other services to clients.

Because the AI landscape constantly and rapidly changes, this is a “living statement”, and I’ll commit to updating it regularly.

Assistive AI

I recognise the difference between assistive AI and generative AI. For many years, I’ve used, and continue to use, assistive AI tools as part of the “mechanical checks” stage of my editing and proofing. This stage takes place after I’ve edited or proofed a document myself and does not replace my own editing. Additionally, in any instance when an assistive AI tool recommends a change to a piece of text I have edited, I use my professional judgement to decide whether or not to accept that change.

Assistive AI tools I regularly use by default include:

Microsoft Word’s “Read Aloud” function to read the text on screen back to me, so I can hear what’s on my screen and catch mistakes I might otherwise miss
Intelligent Editing’s PerfectIt software to check for matters of style and consistency across a longer document
Hemingway Editor’s readability checker to check the reading grade level if a client requires their text to be at a certain grade (note that I use this service without using their built-in generative AI tools).

If you prefer me not to use any of these tools in the course of providing services, please let me know and I’ll re-scope our project to respect your wishes.

Generative AI

I recognise that generative AI (gen-AI) tools such as Claude, ChatGPT, Gemini, CoPilot and similar Large Language Models have the potential to greatly increase work efficiency. However, I also recognise that they currently have significant limits in what they can do well, and that these limits are constantly changing.

Beyond these limits, gen-AI tools can present considerable ethical concerns, including environmental and data privacy issues, as well as the danger of hallucinated/ confabulated information and built-in biases to data. And further, I recognise that gen-AI tools are simply tools – they’re not solutions in and of themselves, and they’re only as good as the human operating them.

Therefore, in recognition of both the potential benefits and risks of gen-AI, I will:

do my best to stay up-to-date on changes in the AI landscape, including changes in the benefits AI confers, and in the potential risks and costs of those benefits
responsibly consider any situations in which I think gen-AI tools might add genuine value to my work for a client (for example, during the brainstorming phase, or for information analysis or synthesis), and carefully weigh the likely benefits against likely risks as I understand them at the time
proactively ask clients whose work I think AI might add value to if they’d like me to use gen-AI tools for that work, then wait for their written permission to use any tools before proceeding
continue to apply my professional judgement to any information a gen-AI tool outputs, including manually cross-checking and validating that information where necessary.

Similarly, I will never:

use gen-AI tools with a client’s text in any way without first asking for and obtaining their written permission to use it in that specific way
provide the direct output that a gen-AI tool generates to the client verbatim without first applying “human intelligence” to validate and edit it to better meet the client’s goals
treat gen-AI as the overall solution for any given project, rather than as one tool to be used at one point in the process of providing my services.

Personal use of AI

While my personal use of AI won’t directly affect a client, it may play a role in whether or not they’re comfortable working with me. That means it’s important to me to be transparent about my use.

I’m currently comfortable with using:

a voice-activated AI assistant to help my neurodivergent brain stay on top of daily tasks (for example, setting alarms and reminders, maintaining shopping lists or checking weather forecasts)
a task management and scheduling system that uses AI technology to support my productivity
an LLM to help me research and outline some of my own blog posts (although my experience is mostly that the LLM’s output acts as a prompt for my brain, and I go off completely in my own direction as I validate the LLM’s information, and then structure and write the post myself)
an AI bot as an artificial “co-worker” when human co-working partners aren’t available to help me stay focused and accountable during work blocks.

I’m not comfortable with, and will not knowingly buy:

books or other written products produced solely by gen-AI
artwork or music produced solely by gen-AI.

I haven’t yet decided where I fall when it comes to products where a human author, artist or musician has transparently chosen to use gen-AI for some part of their creative process.

Where a client has used gen-AI to produce their work

I recognise that many editors have a blanket policy of refusing to edit text for which a client has used gen-AI at any stage. I understand and respect this stance, but it doesn’t reflect my personal position.

Rather, I’ll consider every project for fit with my skills and values on its own individual merits. In particular, I recognise that:

While AI-produced writing can look polished on the surface, it often has significant issues deeper down. These issues can take far more time (and therefore cost more) to identify and fix than most clients realise. In many cases, fixing them can be more expensive than rewriting the piece from the ground up.
In some situations, gen-AI use may go against the policies of the institution for which the text was produced. For example, some universities and many publishers have policies that explicitly prohibit gen-AI use in writing. I will never knowingly provide editing services that help a client “get around” such policies.
Similarly, I acknowledge that copyright law as it relates to gen-AI-produced text is currently murky. If a client discloses previous AI use in a project that they may later want to seek copyright for, I’ll strongly recommend they speak with a qualified intellectual property lawyer before I start any work on that project.

Statement review and further information

This position statement is a living document. The current wording reflects my thinking on AI use as at October 2026.

However, because the AI landscape shifts so rapidly, I’ll actively review this statement at least monthly, and update it if needed.

If you have any questions about my position on AI use, please contact me.