Fred Graver on Using AI as a Writers’ Room, Not a Ghostwriter

Every’s Dan Shipper introduces Natalia Quintero’s conversation with writer and producer Fred Graver. Graver argues that AI can support creative work but that writers need their own point of view; this is his creative philosophy, not a settled test of what AI can or cannot produce.

Graver began at National Lampoon with typing and copy-editing skills. He says understanding when an apparent grammatical mistake was actually a joke helped him contribute as an editor and writer.

He recalls working at MTV Networks during the early web and later joining Twitter. At Microsoft he became a conversation designer, scripting bots and thinking about how people would interact with them.

Quintero asks what linked these career pivots. Graver describes a recurring desire to get involved when new technology and communities appear, rather than treating curiosity as a matter of age.

He recalls Disney Imagineers examining what emerging technology meant for their work and the tension created when audiences gain new access to creative material.

At Best Week Ever, a Tumblr initially created for producers unexpectedly attracted thousands of readers. Graver says the experience taught him to notice audience signals and open the process to them.

For writers using AI, he suggests a document explaining who they are: inspirations, reading, aspirations and tastes. A separate document can describe the project and the question the writer wants to explore.

He recommends asking AI for feedback on the relationship between the writer’s identity and the idea, and treating it as a virtual writers’ room. In his account, the role is to help him think, not simply to write the piece for him.

Graver describes a shared core of entertainment expertise and personal working culture. He wants a system to reflect both how a production is made and the values and preferences of the people making it.

He recalls disputes and friction in conventional productions, then contrasts an assembly-line process with a loop that brings writers, production and marketing into conversation earlier. This is a proposal for organizing work, not evidence that every conflict can be automated away.

Graver divides the work into three broad phases: ideation and design, production, and connecting the work with an audience. He says each phase can inform the others rather than being handed off only once.

For production he gives the example of an AI-animated series that still hires writers. The point is that the mix of human and automated work can differ between a conventional production, a hybrid and an AI-heavy one.

He describes marketing teams testing trailers and adjusting campaigns based on audience response. Bringing that perspective into creative decisions sooner is part of his proposed feedback loop.

The speakers acknowledge that generic AI output is easy to produce. Graver sees structured workflows as a way to keep the work connected to an actual creative aim, while quality still depends on judgment.

Before opening an AI tool, he recommends thinking deeply and getting one’s own ideas out. He relates this to the focused state a writer or artist enters when composing.

Once those notes exist, he brings them into a conversation with a model: who am I, what am I trying to do, and what does this idea need? He treats the model as a sounding board for a process he has already started.

Graver compares this preparation with pitching a project under severe time constraints. A clear understanding of the idea helps a writer decide what to keep and what to cut.

He argues that a creator needs something to say; in his view, asking a model to supply the point of view risks a generic result. That is his assessment of creative practice, and other artists may use the tools differently.

Graver imagines AI helping a writers’ room track a season arc and ask what a character might do. The team still has to judge the ideas and build characters that reflect the problems it wants to explore.

He offers a show-business principle: success can mean understanding what an audience wants before the audience can articulate it. He also values the unexpected personal contributions that emerge when people collaborate.

In his daily workflow he uses custom reader skills and asks Claude whether a draft sounds like him. He may hold a long discussion about whether an idea is worth pursuing and what subsequent drafts need.

He also uses the system to remember a project’s stated direction and to apply a story framework. He describes trying other tools, including NotebookLM, to hear his material in another form.

The conversation ends with a tool that guides the writer through a logline and characters in stages. Graver invites feedback on it and emphasizes continued experimentation.