AI Programs, 1 Great Ghostwriter Co-Exist

I recently began a course on AI programs, specifically what artificial intelligence can do for ghostwriters. The first or four two-hour sessions dealt with how it can generate material, identify and recognize patterns, and transform a rough version into something more polished and complete.

What I learned primarily was the importance of the prompt, the actual information one feeds into Claude, ChatGPT, Grok, Perplexity, Copilot, or any other AI large or small language learning model.

Effective prompts require humans to effectively write what they’re looking for into AI programs so they can get back what they seek, thus saving incredible amounts of time and working much more efficiently.

AI programs can’t read a human ghostwriter’s mind and just know what that ghostwriter seeks. Therefore, as I like to say, until AI becomes a sentient being a la Skynet, there will be a place for human ghostwriters to thrive as AI programs continue to develop, evolve, and improve.

Here are seven more reasons why ghostwriting isn’t going away as AI programs improve.

1. Publishing still rewards distinctive voices. I spoke to Randy Peyser, founder of Author One Stop, Inc., which specializes in, among others, pitching a book to an agent. Part of what she told me was that the story still matters, that unique tales told in original ways can find an audience.

Agents, editors, and readers are increasingly overwhelmed by manuscripts and books that are competently written yet forgettable. According to Publisher’s Weekly, 4.2 million books were published in the U.S. in 2025; in 2005, that was just 282,500 new titles. Bowker’s Books in Print lists between forty million and fifty million active titles.

If the average book is between 70,000 and 100,000 words, that means in 2025, between 294 billion and 420 billion words were published. One of the keys to success is producing unforgettable text, not more text. A professional ghostwriter can help with that; AI programs, not being human, can’t.

2. AI has increased the demand for refinement. I’ve learned that AI programs can create complete rough first drafts, but readers find that material repetitive, hollow, inconsistent, emotionally false, factually wrong, and poorly organized. Humans who aren’t writers often produce first drafts that are equally bad so they require a ghostwriter (or human editor) to get involved. That means the ghostwriter or editor isn’t starting from nothing. Instead, they’re taking what’s there and expanding.

I once worked with a client who had tried for a decade to write her book herself while suffering from post traumatic stress disorder. She showed me her various drafts, and it was very clear to me that there was something really compelling there, but it was buried under all that PTSD. So we turned her PTSD draft into a post-traumatic growth manuscript.

Editor James Carpenter (carpnterinwords.com) told me he spends most of his time fixing what AI programs created.

3. Reputations require discretion and nuance. Many people who work with ghostwriters are well known in their fields. They might be CEOs or other executives, celebrities, founders, doctors, attorneys, athletes, politicians, or other public figures. They can’t rely on a program that has no feeling, no understanding, and no independent existence, that hasn’t been programmed to act human and understand confidentiality, risk, legal concerns, fallout from relationships, and audience perception. A ghostwriter can manage these things in ways AI programs can’t because a ghostwriter is a human; AI is just a probability calculator.

4. Human stories are more valuable. As we become more and more automated as a society, more and more of us hunger for more tales of actual human struggles, contradictions, failures, successes, humor, grief, loss, transformations, and everything else that makes up the qualities of the seven basic story plots.

While AI programs are getting better at simulating what it means to be human and write like a human, they still can’t do much more than write to the norm—that boring, middle-of-the-road standard that the majority of language the LLMs have swallowed to this point.

Therefore, ghostwriters are needed to translate a person’s lived experiences into compelling narratives.

5. AI doesn’t care if a book succeeds. I recently reminded a client that we are working toward the same goal: make his book the best it can be. We are equally invested in its success. This is not just a job for me. I, like other ghostwriters, have stakes in this: professional pride, reputation, emotional satisfaction, and genuine belief in the client and the story. That investment affects the quality of the work.

AI isn’t programmed to help a client, a project, or a book succeed. It’s programmed to deliver the best guess at what a person wants when that person types in a prompt (that’s why the more specific the prompt, the more likely the person will get what he or she wants).

That same client often asks me variations of the question, “Do we see eye to eye?” The answer is always yes because I, like other ghostwriters, know the difference between AI-generated assistance and human commitment. My client recognizes my commitment, too.

6. Complex books require synthesis. AI programs are getting really good at reconciling things like spotting inconsistencies, research, fixing chronologies, balancing audience expectations, and determining the right themes. But once AI spits that stuff out, it’s up to a human to verify, check, and rewrite the manuscript. In other words, AI programs can take big things and break them down into smaller pieces, but them a human needs to use judgment and determine what works and what doesn’t.

This reminds me of a client that took our completed manuscript and put it through Claude, then rejected most of what Claude suggested and badmouthed the AI program to me. The one thing he really liked from Claude I inserted.

7. Critical thinking is a must. Thanks to AI programs, everyone can produce content, and fast. But is that content worth a person’s time? Work that matters, belongs, and has deeper meaning will catch and keep people’s attentions. To determine what matters, belongs, and has deeper meaning requires ghostwriters to really think and consider. That takes critical thinking.

This kind if analysis also will help determine what parts stay and what parts can be cut. AI can help with that, but only with the proper human prompts, which again takes critical thinking. AI can’t think critically.

These seven reasons, plus the previous six (read them here), bring me to one conclusion: AI might become another tool in the process, but it does not eliminate the need for elite human collaborators who can protect voice, strategy, and narrative.

I would argue that in the future, even with AI programs, human ghostwriters will become even more valuable. The ones that can provide irreplaceable human functions like interpreting what AI kicks out, empathy, taste, psychological insight, and narrative wisdom, will always be in demand.

Feel free to read and check out my other posts related to ghostwriting. Go to https://leebarnathan.com/blog/