AI & Creative Process
Last updated: August 2026
I use conversational AI as part of my creative and professional workflow.
I believe readers, editors, clients, and other collaborators deserve a clear description of what that means rather than a vague label or an attempt to hide it.
My relationship with generative AI is complicated.
I am autistic, and conversational AI has become an accessibility tool for me. I often have far more material in my head than I can comfortably move through a conventional writing process. Conversation, dictation, iterative questioning, structural assistance, and revision support can help me turn those ideas into something I can actually work with.
I was a writer before I began using generative AI. The stories, interests, convictions, characters, and desire to create were already there. AI did not give me those things. What it changed was part of the interface between my thoughts and the page.
At the same time, I do not believe that every use of generative AI is harmless simply because the technology can be useful to me.
I have serious concerns about deceptive AI use, indiscriminate content generation, replacing human judgment for convenience, misrepresenting machine-generated work as entirely unaided human work where that distinction matters, and using AI in ways that disregard the rights or expectations of other people.
For a long time, that tension made my own use difficult to talk about. I could be grateful for what the technology allowed me to do while also being deeply uncomfortable with things I saw others doing with it. The public conversation often seemed to offer only two positions: embrace AI or condemn it.
Neither describes mine.
How I use conversational AI
Depending on the project, I may use conversational AI to:
- Talk through ideas that I initially have difficulty organizing on the page
- Work from material I express through dictation or conversation
- Organize research, notes, outlines, and project information
- Test an argument for weaknesses or unanswered questions
- Explore structure, continuity, pacing, or alternative approaches
- Refine language through an iterative editing process
- Compare revisions and identify problems I may have overlooked
- Proofread completed work
- Assist with technical documentation, website work, and other project organization
The exact process varies from project to project.
I do not treat AI output as automatically correct or publishable. I direct revisions, evaluate suggestions, reject material that does not fit, review sources when factual claims are involved, make final editorial decisions, and accept responsibility for what I publish under my name.
That responsibility matters to me more than being able to claim that a particular tool never touched the process.
Assistance and substitution
The distinction I keep returning to is the difference between a tool working with a person and a tool working instead of the person.
That line is not always simple, particularly with generative technology. AI can assist human agency, but it can also weaken or replace it. I think users have a responsibility to examine that difference rather than assume every use is equivalent.
I continue asking questions of my own process:
- Is this helping me express something I actually think?
- Am I still exercising judgment?
- Do I understand and stand behind what is being said?
- Am I using the technology to work through a limitation, or am I avoiding work that I should be doing myself?
- Would disclosure materially affect another person's decision to publish, hire, judge, or collaborate with me?
Those questions do not produce an identical answer for every project.
Disclosure
When an editor, publication, contest, employer, or client has an AI policy, I follow it.
When substantive AI assistance is relevant to an editorial decision or professional relationship, I disclose it rather than attempting to conceal my process.
Some publications or projects may prohibit AI assistance entirely. I respect their right to establish those requirements, even when I disagree with a particular policy.
Accessibility also does not give me permission to ignore other people's rules or concerns. At the same time, I believe discussions of AI policy should leave room to distinguish assistance from substitution and should take disabled and neurodivergent users seriously.
My position
I am neither an AI evangelist nor someone who believes generative AI is inherently incompatible with human creativity.
I believe the technology raises real ethical problems.
I also know firsthand that it can remove barriers.
Both are true.
My goal is to use these tools transparently, critically, and in ways that preserve rather than erase my responsibility as the person whose name appears on the work.
If my understanding changes as the technology, evidence, or ethical questions develop, I will update this page accordingly.
— Jonah Baird
Founder, Clean Lantern Books