Is Claude Helping or Hurting My Creative Process?

Artificial Intelligence / By / / Artificial Intelligence, Claude, Knowledge management

Claude now features at most stages of my creative process: is that making me more creative or less?”

I often get ideas while working on another post or walking the dog, and I try to capture them immediately. Blog post ideas go into my list of potential content ideas.

But ideas don’t always arrive that way. Two weeks ago, I published a blog post called A Fork in the Road: Will AI Make Us Smarter, or Leave Some of Us Behind?. The idea that led to that post came up during a discussion with AI. It intrigued me, and it ended up in my list of potential content ideas.

In fact, Claude came up with the title for this post. I had two similar ideas on the creative process, one relating to AI and the other to using my Zettelkasten. My idea was to merge them, and Claude helped me with the layout, including the title and even an alternative.

David Kadavy, in his book Mind Management, Not Time Management, describes the creative process as having four distinct stages:

  1. Preparation
  2. Incubation
  3. Illumination
  4. Verification

I explored these four stages in my blog post Unleashing Your Creative Potential: The Four Stages of Creativity.

Is Claude a creative partner?

Most of my creativity comes from merging ideas, often ones I’ve consumed in other people’s content. I have even written about the key takeaways from a book that has had an impact on me. Large Language Models like Claude can be creative in the same way.

So Claude can be creative. But when it’s working with me, it’s part of a working relationship in which I drive the process, even when I take Claude’s work into account. It’s me who has an objective or goal in mind for what we’re working on.

In my post What impact would Artificial Intelligence have on the Knowledge Innovation loop?, I explored AI’s role in invention, which is a form of creativity.

Preparation: where Claude earns its keep

When I first started blogging, there was no AI to help me, so I had to rely on what I could remember and the links to related notes in my Zettelkasten. These links in themselves can stimulate your creativity. But there is one limitation: the starting point depends on my own memory.

So what does Claude bring to the table? Firstly, it’s a collaborative partner that can find its way around my vault quickly and easily. It suggests permanent notes I had forgotten about, but which were still part of the web of ideas I had created with my Zettelkasten.

Over the last few months, probably since the release of Claude Opus 5, it has also started to warn me when a writing idea might overlap with, or even take traffic away from, another post. And it flags when an idea has already been explored in another post, so I might want to consider linking to it.

Again, this is something I would have thought about, but I would have missed some of them.

But most importantly, using Claude in this way doesn’t cross any of my red lines, as the ideas it puts together are just a framework to help me start writing. What I write is completely down to me.

A screenshot showing my edit of the original AI generated layout to mirror the reality of what had happened during the writing of the post.

Incubation: does Claude leave room for my subconscious?

Incubation is the stage of the creative process when you’re not actively thinking about a problem.

It depends on you having already worked on the problem, or on the ideas it emerges from. For my writing, that could be the books I’m reading, the synthesis of fleeting notes into literature and permanent notes, or the drafting of content such as blog posts and newsletters.

This mixture of ideas needs time to ferment, which I give it during my downtime and sleep. I often get ideas at night or in the bathroom.

I don’t know whether Claude has had any impact on how often I get ideas. But in my experience, the following things do:

  • Free time: I’ve noticed that I get a lot of ideas if I take a break for a few days.
  • Walking the dog: Not having a dog reduced the number of ideas I had. A lot of them came when I walked my old dog, George. As Louis, my new puppy, grows up, I’ve noticed that I’m starting to get ideas when I’m walking him.

My current thinking is that Claude is helping me to mix ideas, and should therefore help with incubation, but I have no evidence to back this up.

As I have breaks from Claude just like every other colleague or tool, I still think my subconscious has space to work.

Illumination: whose insight is it?

Illumination is the part of the creative process when you get a fresh insight. It can be the new idea that comes to me when I’m walking the dog, reading a book or linking permanent notes in my Zettelkasten, or, increasingly, during a discussion with Claude.

The question of whose insight it is could apply to any of them, as most insights come from a combination of linked ideas. But I would argue that an insight is personal to the person who has it, so my insights are personal to me.

Even if an insight emerges from a conversation with AI, I treat that conversation as a source, as the insight arose from something that can be referenced. But I got the insight, and that means it’s mine.

The best example I have for Illumination is my post A Fork in the Road: Will AI Make Us Smarter, or Leave Some of Us Behind?, which I mentioned in this post’s introduction.

The idea for that article came from a discussion I had with Claude about an idea I’d had while writing. During this conversation, Claude suggested an alternative to what I had been thinking, and that idea resonated with me. I had been thinking of it as an either/or situation, but Claude asked, “Why can’t it be both?”, and put forward an argument to back up its thinking.

Verification: a second pair of eyes

This is where we all need to be careful with AI. All models are trained to please the user, which means they tend to confirm your beliefs, and that could potentially lead to cognitive problems. You can explore these potential problems in my post Cognitive Mirror Syndrome: Should We Be Concerned?.

That’s why I’d urge you to include something like “Give honest and constructive feedback” in the system prompt of whichever model you use. It’s what I use. I want Claude to help me improve, not stroke my ego.

Claude helps me a lot here by editing my posts, which is something AI has always been quite good at. I used to subscribe to Grammarly, but I cancelled it a couple of years ago because of AI.

Claude Opus 5.5 is a really good editor. It fixes the grammar and spelling errors for me, then lists things I should reconsider, with suggestions on how to improve the quality of the post. More often than not, the suggested changes are worth making. Again, I make these changes myself, so they remain in my words.

So, is Claude making me less creative?

This is the big question, isn’t it? Is Claude making me less creative?

In general, I think the answer is no, but I do think my red lines play a part in this. Both Claude and I respect those red lines, and they force me to engage with ideas, whether they’re my own or other people’s. You can learn more about my red lines in my post Introduction to my AI Knowledge Framework level 0.

The work I’m offloading to Claude is either administrative work or things that are of no interest to me, which frees up time for more demanding work. Read my post Cognitive Debt: The Real Cost of Offloading to AI to explore the risk further.

I also think Claude is helping me to find new perspectives and angles, even if these might pull me towards the average of human thought. This averaging effect comes from AI being trained on much of written human knowledge. I explored this further in my post Using Claude as a reading partner.

Why don’t you join my monthly newsletter and keep up to date with our most recent content?

Further reading

Leave a Comment

Your email address will not be published. Required fields are marked *