Modal Systems

Russell T Davies and the Writer Who Was Always There

The writer did not disappear.

A television writer has asked what television writers are paid to do if artificial intelligence produces their first drafts.

This is going to get awkward.

On September 29, BBC director-general Matt Brittin told an audience at the Royal Television Society convention that somebody had sent him a fifty-minute, AI-generated episode of Doctor Who.

He watched it.
He said he thought it was “pretty good.”

And the British television industry has apparently spent the following two weeks trying to recover from this unimaginable breach of professional conduct.

Brittin said something nice about a television program.

The Doctor Regenerates Into an Employment Dispute.

On October 8, Deadline reported that BBC employees had challenged Brittin during an internal staff meeting.

One employee raised a serious concern about Doctor Who being produced in Cardiff and the BBC's responsibility to support creative industries throughout Britain.

Good. That is a real institutional question.

Doctor Who has supported a production ecology in Wales: performers, writers, technicians, craftspeople, facilities, suppliers, and careers that only exist because shows are physically made somewhere by people. A change in production methods could change that.

The BBC is also facing substantial financial pressure and planned job reductions. So workers have every reason to ask exactly what management intends to do, which functions it proposes to automate, and who will bear the cost.

Brittin's enthusiasm matters.

His procurement decisions, commissioning decisions, staffing decisions, and obligations to the people making BBC programmes matter much more.

The director-general thinking a fan film was pretty good has not itself eliminated a single production job.

It does not establish that the BBC plans to commission synthetic replacements for its actors or writers.

And, critically, nobody has established that this film actually was pretty good.

Brittin offered an aesthetic judgment. It received approximately two weeks of institutional weather.

Then Russell T Davies arrived.

And somehow made the argument even stranger.

Russell, What Is a Writer?

On October 14, Davies criticised Brittin's remarks as clumsy, warning that writers and artists face some of the earliest consequences of AI adoption.

That is a defensible objection to the judgment of a director-general speaking during a period of workforce anxiety.

Davies has spent decades making television. He understands that creative professions are sustained by institutions, commissions, training, collaboration, money, and opportunities to work.

Then he was asked about screenwriters using AI to prepare first drafts.

His response:

“What are you paid for otherwise? What are you there to do?”

Russell?

To write the fucking television program.

What do you think happens after the first draft?

A screenwriter develops dramatic structure, characters, motivations, scenes, dialogue, pacing, relationships, and thematic coherence.

A screenwriter rewrites.

A screenwriter discovers that the ending does not work.

A screenwriter removes a character who was consuming twelve minutes of television without contributing anything to the story.

A screenwriter changes the entire second act because a better dramatic relationship emerged during revision.

A screenwriter may have to redesign scenes around available locations, casting decisions, production budgets, performance, and the material constraints of filming.

Some writers develop their stories through the first draft itself. That discovery process can be indispensable to their work. Other writers plan exhaustively before writing a scene. Others collaborate. Others produce extensive treatments, outlines, and dialogue experiments before arriving at a complete screenplay.

These are just different workflows.

The first draft is one important stage in a larger creative process.

If a writer uses a language model to prepare an initial version, the existence of that document tells us astonishingly little about the human work required to produce the final screenplay.

Perhaps the writer contributes almost nothing afterward.
Perhaps they transform the entire thing.
Perhaps the model produces such terrible material that the writer wastes three days repairing a problem they could have avoided entirely by just writing the goddamn scene themselves.

Great. Now we can examine the work.

Davies has mistaken a production operation for the whole professional function.

And he has done this while defending the profession.

There Was a Writer in the AI TARDIS.

Now the investigation becomes particularly funny.

Brittin has not publicly identified the episode he watched.

That definitely matters. We cannot establish which production he praised or how good it was.

But there is a prominent candidate.

In September, an unofficial fan production called The Curse of Kilmorag appeared online as part of The Unmade Season.

It ran approximately fifty-five minutes.

It attracted hundreds of thousands of views.

It used generative video technology to produce most of its footage, including synthetic performers.

And Doctor Who TV documented something rather important.

The screenplay was written by a human being.

His name is Thomas Vittlowe.

The production's own trailer description identifies him as a screenwriter and describes approximately two months of production and editing.

The principal video tools included LTX 2.5 and MiniMax.

So, this was not a fifty-five-minute television drama appearing spontaneously when somebody typed DOCTOR WHO GOOD EPISODE into a machine.

A human conceived and wrote a story.
A human organized a production.
Human editorial decisions assembled generated material into a continuous dramatic work.

Other creatives also assisted the project, according to Vittlowe's account of its production. The production used artificial intelligence extensively.

Both descriptions can be true.

We do not know whether this was Brittin's pretty good episode. The runtime is close, and the timing fits, but neither fact establishes identity.

We also do not know whether the finished work was pretty good. Contemporary reactions were divided. Some viewers liked the writing; others criticized the synthetic performances and visual discontinuities.

That is television criticism.

Go criticize it.

Also, the film reportedly reproduced David Tennant's likeness and voice without his permission. So that raises a separate, substantial question about performers' rights, consent, and the use of recognizable identities.

Go investigate that too.

But notice what happened to the human author when people began describing the project as AI-made.

He disappeared.

There was a writer.
There was a production history.
There were editorial decisions.
And we replaced all of it with the name of one instrument.

The writer was there the whole time.

The First Draft Was Human.

One more distinction before everybody begins celebrating their preferred side.

The Curse of Kilmorag is not evidence that AI can already write excellent first drafts. The available production account says its screenplay was hand-written.

It is evidence of a different problem: how readily the description AI-generated can erase human authorship and labor elsewhere in the production.

Davies's first-draft objection remains a separate argument. And there is a stronger version of it.

Suppose a broadcaster previously paid junior writers to draft episodes, develop treatments, test dialogue, and learn the craft through supervised assignments.

Management replaces much of that work with generated drafts.
Senior writers remain employed to polish the outputs.
The broadcaster saves money.
The junior writers lose paid opportunities to develop.

Five years later, the institution discovers that the experienced writers it relied upon are retiring and the people who should have succeeded them never acquired the same opportunities to practice.

There is a real structural problem.

The institution has preserved its immediate production output while damaging the field through which future writers are made.

That is not a speculative moral category. That is a concrete mechanism whose actual occurrence can be investigated through commissions, contracts, staffing, career progression, and the distribution of writing assignments.

It is also completely different from a writer voluntarily using a tool inside their own process.

A broadcaster replacing a paid writing position and a writer experimenting with a drafting instrument do not describe the same transition.

The contractual terms are relevant as well. A writer commissioned specifically to deliver original human-written material may violate that agreement by substituting generated text. A writer commissioned to develop and deliver a finished screenplay under different terms may retain substantial creative responsibility while using assisted drafting.

We need to know which arrangement exists. Otherwise, the first draft has become a purity test for the profession, and the actual labor relation has disappeared behind it.

Matt Is Not Entirely Off the Hook.

Brittin also suggested that billions of people now have tools capable of producing content at Hollywood-studio standards.

That's an extraordinary claim.

Where is the standard?

Where is the comparison?

A convincing shot is not a complete film. A complete film is not necessarily a good one. A synthetic actor who looks persuasive for twelve seconds has not demonstrated the expressive range, direction, continuity, or performance demands of a television series.

A generated image of a studio production does not automatically reproduce the competence of the studio.

And a useful new production instrument does not establish that the institution employing it can dispense with the professional ecology that formerly supplied its output.

If the BBC intends to use generative systems, it needs to know which functions those systems actually perform, which people continue to perform the others, and what responsibilities remain attached to the resulting work.

Its public-service obligations do not vanish because one impressive demonstration arrived by email.

Brittin can think a film is good. He still has to govern the broadcaster.

Davies can think AI-assisted screenwriting is artistically terrible. He still has to distinguish the artistic judgment from a claim about employment, authorship, or professional legitimacy.

Both have substantial responsibilities here. Neither gets to settle the whole subject by reacting to a demonstration.

I Already Did This With the Gamers.

On October 3, Modal Path Ethics published This Is Not Worker Solidarity You Morons.

The proposition was simple.

If a developer uses generative AI, identify what the use actually did.

Did it worsen the work? Displace somebody? Violate a right? Destroy a legitimate production relationship? Create a new dependency? Transfer the benefits of creative labor away from the people performing it?

Those questions remain available when the developer is making television instead of videogames.

And they become more urgent when the institution is a public broadcaster with enormous cultural influence and obligations to an existing workforce.

But there is another side to the transition.

Vittlowe described himself as somebody who had been trying to enter the television industry for years.

Generative tools helped him attempt a production at a scale he could not readily achieve through conventional financing.

That counts too.

A person who lacks a studio budget may gain the ability to make a film.
A person who has spent decades building a career inside a studio may face new pressures as the same technology changes production economics.
A performer may find their likeness replicated without consent.
A broadcaster may retain output while losing its apprenticeship pipeline.
An audience may receive something new, terrible, or some appalling combination of the two.

These are different loci, relations, and transitions.

The possibility of one gain does not erase another person's loss.

The possibility of another person's loss does not establish that every gain is illegitimate.

And the person who used the machine does not stop being a person.

The Writer Is Still There.

Russell T Davies is entitled to despise AI-written dialogue and write every word himself.

He is entitled to argue for production contracts that protect human authorship, junior writing opportunities, and creative work against exploitative substitution.

He may have excellent reasons to believe that writing an initial draft is essential to his own creative process.

A screenwriter who uses AI does not automatically cease being a screenwriter, any more than a filmmaker who uses digital compositing ceases being a filmmaker.

The instrument does not decide who authored the work.

The actual production relation does.

And if the machine really produced a pretty good first draft, the question remains what the human writer did with it next.

Perhaps nothing of consequence.
Perhaps enough to make the result extraordinary.
Perhaps the finished work would have been better off without the instrument.

We have to look to learn.

Meanwhile, the BBC must answer its workers' legitimate questions about employment, rights, regional production, and the future of the industry's training and creative infrastructure.

Those questions do not require Brittin to pretend a film he saw was bad, actually.

They do require him to distinguish a technology demonstration from a production strategy.

A writer deserves protection against exploitation.
A writer also deserves to be recognized when they use an unfamiliar tool to make something.

That is the same principle.

And if the most prominent candidate for this allegedly AI-made Doctor Who episode really is the one Brittin watched, then an extraordinary amount of public argument about replacing writers has taken place around a production with a human-written screenplay.

The BBC director-general may have praised the machine for work a human performed.

The former showrunner may have defended writers by reducing their profession to the production of initial drafts.

And the person who actually wrote the possible subject of the controversy has spent the entire argument just standing right over there.

The writer did not disappear.

MULTIPLAYER: Abdul El-Sayed and the 50/50

IGN Has Mistaken a Cool Crocodile for an Argument

Applied Case: Claude Saw Someone Matching the Description