Preface: Re-teaching myself the new filmmaking pipeline, now with AI (netflix please notice me). co-written with Claude. NOW WITH AUDIO.

Even after a film goes into production — after shooting begins — rewriting continues. On-set rewrites are a daily feature of most productions. Scenes are revised overnight to address problems discovered during the day's shooting. Dialogue is adjusted because an actor finds a line unplayable. A location becomes unavailable and scenes must be restructured around what is accessible. An actor's performance reveals a dimension of a character not anticipated in the script, and the writer — if they are present on set — adjusts subsequent scenes to develop it. Some productions have the original writer on set throughout shooting. Many do not. It is entirely possible for a writer to deliver a script, receive no further involvement, and then watch the finished film discover that their dialogue has been changed, their scenes restructured, their characters fundamentally altered.

Even after shooting is complete, writing continues in a form. The editing of a film is a fundamentally narrative process — editors and directors assemble the footage into a story, and that assembly process involves decisions that are essentially writing decisions: what scenes to include or cut, in what order, with what emphasis. A film can be significantly changed in the editing room — characters can be reduced or expanded, storylines can be dropped, the tone can shift dramatically from what the script anticipated. Some films have scenes written and shot specifically to address problems discovered in editing — reshoots that might involve new scenes being written long after the original writing process was considered complete. The total writing process for a major studio film, from initial commission to final cut, typically spans two to five years. The number of drafts — counting all writers and all stages — might reach twenty to forty or more on a complicated production.

One of the WGA's most important functions is credit arbitration — the process by which the Guild determines who receives on-screen writing credit for a film or television episode. Credit arbitration exists because, as we discussed, major productions often involve many writers. The studio and producers submit a list of all writers who contributed to the final script, along with all draft materials. A panel of experienced WGA members — who are not identified to the parties involved, to prevent lobbying — reads all the materials and determines which writers contributed sufficiently to the final script to merit credit. The WGA's credit rules require that a writer contributed a substantial portion of the final script — typically at least 33% of an original screenplay or 50% of an adaptation — to receive credit. The rules are designed to protect original writers from being displaced from credit by subsequent rewriters, and to ensure that credit reflects genuine creative contribution rather than political influence. Credit arbitration is frequently contentious. Writers who believe they have been denied credit they deserve can appeal. Studios and producers sometimes lobby — in violation of WGA rules — for credit outcomes that serve their interests. The process is imperfect but represents the most systematic attempt in the industry to ensure that writing credit reflects reality.

Screen credit is not merely a matter of prestige — it has direct financial consequences. Residual payments are tied to credit. Future employment opportunities are significantly shaped by credit — a writer with a produced credit on a successful film commands considerably more in their next negotiation than one without. Award eligibility is determined by credit. The entire economic architecture of a screenwriting career is built on the foundation of credited work. This is why credit arbitration is so important and so fiercely contested. It is not a vanity dispute — it is a dispute about economic justice and career survival.

The WGA strike of 2023 was the most significant labor action in Hollywood since the 1988 strike and one of the most consequential in the history of the screen industries. It lasted 148 days — from May 2 to September 27 — and was joined midway through by the actors' union SAG-AFTRA, effectively shutting down film and television production across the industry. As discussed, the shift to streaming had significantly reduced writers' residual income. The Guild sought a formula that would restore meaningful residual payments for streaming content. The streaming era had produced a troubling trend toward mini-rooms: very small writers' rooms, sometimes of only two or three writers, working for short periods, without the traditional staff writer positions that had historically provided entry-level employment and on-the-job training. The Guild sought minimum staffing requirements that would preserve the writers' room as a functional creative and employment institution. The studios sought the right to use artificial intelligence in the development process in ways that could potentially displace writers — generating first draft material using AI systems and then hiring writers at lower rates to polish AI-generated content. The Guild sought explicit protections against this use of AI, including provisions that AI-generated material could not be used as source material for WGA-covered work and that writers could not be required to work on AI-generated scripts. The strike was ultimately resolved in late September 2023 with a contract that the WGA leadership described as exceptional — securing significant improvements in streaming residuals, minimum room staffing requirements, and AI protections. Whether the protections are sufficient to address the actual threat AI poses to the profession remains to be seen.

The AI question is the most urgent and most uncertain issue facing screenwriting today. It deserves careful, honest treatment — because both the anxiety and the optimism surrounding it contain genuine substance. Current large language models — GPT-4, Claude, Gemini, and their successors — can generate screenplay-formatted text of considerable technical fluency. They can produce scene descriptions, dialogue, plot outlines, and character descriptions that are grammatically correct, structurally coherent, and sometimes surprisingly plausible on a surface level. They can also perform specific, bounded tasks with genuine usefulness — generating multiple versions of a scene for a writer to react to, producing quick first drafts of expository dialogue, suggesting structural options for a story problem, analyzing a script for pacing and structural issues.

The limitations of current AI systems in creative writing are real and significant — though it is important to acknowledge that they may not remain permanent limitations. Current AI systems have no genuine experience of being human. They have processed enormous quantities of text about human experience but they have not lived — have not loved, grieved, failed, been humiliated, been surprised by beauty, been devastated by loss. The specificity of genuine human experience — the particular texture of a specific life that makes a character feel real rather than generic — is not something AI can draw on because it has no life to draw from. Current AI systems tend toward the generic and the average. They are trained on enormous quantities of existing text, which means they reproduce patterns — genre conventions, familiar character types, predictable emotional beats — with high efficiency and low originality. The quality that distinguishes great screenwriting from competent screenwriting — the unexpected observation, the character detail that feels simultaneously surprising and inevitable, the structural choice that defies expectation while satisfying a deeper need — is precisely what AI systems are worst at producing. Current AI systems have no aesthetic judgment. They cannot evaluate their own output with genuine discernment — cannot recognize when something is working at the level of genuine drama versus merely functioning at the level of structural competence.

The artistic limitations of current AI, however, may not fully protect screenwriters from its economic impact. The threat is not that AI will replace great screenwriting — it is that it will replace adequate screenwriting, and that the industry will decide adequate is sufficient for many purposes. A studio developing a high-concept franchise film does not necessarily need brilliant, original dialogue — it needs competent, functional dialogue delivered quickly and cheaply. If AI can produce serviceable first drafts that human writers then polish, the economics of development change significantly. The number of drafts that require human writers at full rates decreases. The demand for mid-level writers — not the showrunners and the script doctors at the top of the market, but the working writers who make up the bulk of the profession — potentially shrinks dramatically. This is the real concern behind the WGA's AI provisions in the 2023 contract. Not that AI will write Chinatown — it won't — but that it will make human writers economically marginal in the development process for the large middle of the market that is neither the prestige end nor the purely formulaic end.

Beyond the economic question there is a deeper cultural question that the AI debate in screenwriting raises. If films and television shows are written — even in part, even in first draft — by AI systems, what is the relationship between those works and human experience? Storytelling has always been a way that human beings process, share, and make sense of their experience. A story told by a system that has no experience is a fundamentally different kind of cultural object — not necessarily worthless, but different in its origins and potentially in its resonance. Audiences may or may not care. Commercial entertainment has always been partly formulaic — the studio system's production line approach to storytelling was not so different in principle from using AI to generate competent genre material. But there is something worth preserving in the idea that the films and television that shape how we see ourselves and each other are made by people who have lived the experiences they are dramatizing. This is not a new tension — it is a version of the tension between commercial entertainment and artistic expression that has run through the entire history of screenwriting since the nickelodeon era. But AI makes it more acute and more urgent than it has ever been before.

The profession of screenwriting in 2026 is simultaneously more visible, more discussed, more culturally respected, and more economically precarious than at any previous point in its history. More visible because the prestige television era has made showrunners public figures in a way that film writers almost never were. More discussed because the WGA strike brought the realities of the profession to public attention in a way that labor disputes rarely achieve. More culturally respected because the critical conversation around prestige television regularly treats the writer's vision as the primary creative force — the showrunner as auteur. More economically precarious because the streaming revolution disrupted the residual structures that sustained mid-career writers, because mini-rooms replaced full staffs, because the spec script market never fully recovered from the 2008 financial crisis, and because AI looms as a potential structural threat to employment across the market.

The writers who are thriving are those at the top of the market — the showrunners, the established feature writers with track records of commercial success, the script doctors whose diagnostic skills cannot be replicated. The writers who are struggling are those in the middle — the working professionals who have credits and experience but have not yet broken through to the level where their position is secure. The 2023 strike won important protections. Whether those protections will be sufficient as AI capabilities continue to develop, as the streaming landscape continues to consolidate, and as the economics of content production continue to evolve — that remains genuinely uncertain. What is not uncertain is this: the need for genuine human storytelling — for writers who have lived, who have felt, who have something specific and true to say about what it means to be alive — is not going away. Every generation discovers again that the stories that matter most are the ones that feel most urgently human. And the most urgently human stories require urgently human writers to tell them. More on this in the next post. ☀️