Future Magic (IV. Drama as a Writer’s Medium)

(Note: this essay is the final in a series.)

The main promise of AI imaging is that the writer will - or at least could - become the general creator; the author not just of the words, but of the work. A drama, even one that appears to involve hundreds of crew and dozens of actors, might in fact be entirely the creation of one person, using sophisticated technology to, as the old phrase goes, “realise his creative vision”.

Of course, this does not mean that dramas will necessarily be better, but it will revolutionise what they are.

---

The BBC started as an outfit working purely with radio. Once television became feasible it branched out into that. For this purpose, the BBC acquired small spaces all around London and converted them into television studios. These were distinct from film studios in that the cameras were not discrete machines chewing through celluloid, but wholly electronic and hooked into a complex system for live editing and broadcast. After a decade of being scattered across the city, BBC television operations moved into a massive new headquarters, BBC Television Centre. This was a building custom-designed for producing high volumes of television in every genre. It housed every department needed for pre-production, production and post-production.

Around 2010 the BBC began to retire Television Centre. I read some industry experts discussing how it had become redundant. The main reason astonished me: the entirety of post-production - foley, grading, editing, titling, graphics, visual effects, sound effects, storage - could now be done by a single person “using an averagely powerful laptop”. Advancing technology had rendered superfluous all of these different departments, and all of the specially-trained people staffing them along with the many specialised tools they used. All the functions of the post-production phase could now be packed into a simple laptop and performed by one person, even in his home.

Post-production had become the work of one person, or a tiny number of people, doing everything on an extremely versatile yet off-the-shelf device.

Of course, the production phase could not be so reduced. You still needed actors performing, on a set built by people, wearing costumes made by people, shot with cameras operated by people, lit with lights controlled by people, and all of it directed by someone. These many jobs still had to be done “in meat-space”, by humans.

The pre-production phase was likewise undefeated by technology. You still needed people to scout for locations, design the studio sets, design the costumes, assemble the crew, audition the actors, do the costing for all of this, find the money for it… and of course, write the script that is the embryo for all of it.

Post-production was usurped by digital tools on a laptop twenty years ago, and I surmise that AI will do the same to production and pre-production. The only job it cannot usurp is the writing of the script. (In theory it could do this too, but I doubt many people would be interested in watching such a thing.)

Cinema was always “a director’s medium”. Due to the sheer number of options available to the director, his choices (and therefore his tastes, personality and concerns) mattered immensely to the finished work and could outweigh the influence of the writer. This was in stark contrast to television drama. Due to limitations of time, money and the multi-camera mode of production, TV directors simply didn’t have many options as to how to shoot a scene (this is why most of them yearned to graduate to cinema). Therefore, the script was all-important. The writer, not the director, was “the author”. TV drama was “a writer’s medium”.

In the 1980s British TV drama production shifted from video to film and from multi-camera to single-camera, generally becoming more cinematic. The TV director now operated much more like a cinema director. By the 1990s the director, not the writer, was “the author”. TV drama shifted its focus from script quality to spectacle, style, aesthetic and tone. In my opinion this was a huge cultural loss for Britain.

For two reasons, I believe AI could restore drama as a writer’s medium:

1. A single person - probably the original writer - can create the entire thing on his own.

2. Hollywood, with all of its special effects and explosions, cannot compete with a 15 year-old with a laptop. He can achieve the very same special effects and explosions using AI, therefore they become worthless as “selling points”. The same applies to having A-list actors - they too can be replicated with AI, if desired. Leaving acting quality aside, the only way for a drama to compete with any other will be the quality of its script.

This becomes more relevant when we take into account that Hollywood has forgotten how to write good scripts. Films today are dreadfully written. The industry has run out of things to say. Or rather, it has become so ossified and formulaic that it no longer has the inclination nor the leeway to say anything true. The reasons for this are outside the scope of this essay, and I don’t think I fully understand them anyway. But it is clear that Hollywood is no longer in the business - if it ever was - of reflecting the human experience honestly. A script from a thoughtful 30 year-old who has lived a bit but is back living with his parents is likely to be more honest and engrossing than anything conjured in Hollywood by a blue-haired moron, overseen by psychopathic moguls, and fine-tuned by a succession of unimaginative focus groups.

But, even in the good old days of Hollywood’s heyday (pick your decade), films were made by consensus. Rarely if ever did a writer’s “vision” get to the screen. Indeed that was not even seen as the goal. It was the director’s vision that was prized, with the writer’s script being merely a guide for him. (As a sidenote, it is striking how many distinctive and creatively-successful films happen to be written and directed by the same person. It’s almost as if something special occurs when a man is allowed to nurture into adulthood the child that he created. If the original writer can get to be the director, this seems to bode well for a film.)

So in theory it was the director’s vision that should get to the screen. However, in practice, it was often the studio’s vision that was realised. At any rate, virtually every director had to compromise with the studio with every film he made. Sometimes this was for the best; directors, like art students, can get lost in pursuing shock and controversy for its own sake, and some directors are simply not very talented and need “the higher-ups” to save them from themselves. But even so, it is the case that film-making, certainly in the mainstream, has always seen the individual voice get tamed or muted, often completely.

We were horrified when O’Brien reminded Winston that, in Oceania, all books are written by committee and “no book is produced individually”. We understand instantly the assault on human dignity and creativity that must happen when an individual voice is crushed – not just by ideology, but by group consensus. We know what it means: a lack of honesty, distinctiveness, originality and truth. If we feel this way about book writing, why not drama production?

---

Note that I say “drama production” - not film-making, cinema or TV. That is because what we are talking about is revolutionary change. The forms of old will become irrelevant, the limitations obsolete. Cinema will be seen as an art form of the 20th Century, just as the “electronic theatre” which I dearly love is now seen as a thing specific to 1970s Britain.

We already see Hollywood slipping into the past. In my adulthood, the annual Oscars has gone from something everyone knew about and many watched, to something that nobody cares about at all. This is for a whole slew of reasons that include Hollywood sleaze and corruption, bad writing, rampant wokery, endless sequels and reboots, and lamentable “cape shit”. With these deficiencies plus the competition from video-gaming, Hollywood as an industry seems to be dying, and nobody is even mourning it.

Having said all of that, the three-act 90-minute format that Hollywood popularised will probably remain, because of audience habituation and simply because it works well for a lot of stories, but I doubt it will remain the default. As cinema-going has been usurped by streaming at home which makes any format possible, the one emerging as the favourite seems to be multi-episode arcs, whether three or twenty episodes, and sometimes multiple seasons of those. This much longer form provides room for characters and worlds to breathe, for plots to have more twists than could possibly fit into 90 minutes, and for the story to have more novelty than is afforded by the three-act structure after a century of familiarity.

Of course, limitations can be inspiring and they can provide the structure that a story needs, so, to repeat, I don’t think we are seeing the death of the “feature film” format; I just think writers won’t be restricted to it unless they choose to be. And good writers willingly choose restrictions, because they know that too much freedom is deadening for creativity. But, in future, this really will be the writer’s choice; he will decide on the story and on the duration it needs to be told well.

---

There are dramas I have had in my head for ten years, twenty years, even more. I will describe them in broad terms.

• a children’s animated show about a community of animals, each one having a personality based on people I knew in my childhood - old Scottish people born in the 1920s, etc. It would be light-hearted and show how different characters get along despite their peccadillos. It would show a community that dealt with eccentricity, class, simple people and intelligent people. It would be an expression of rural Scottishness. At its heart would be a love of home and of community.

• a set of character studies of fictitious left-wing activists circa 2011, immediately post-Occupy

• a sci-fi romance about Britain and being a young man (mostly written in 2012). This is my eternally unfinished novel

• from the above, a spin-off novella about the life of a Brutalist town from its inception to its demolition

• a set of novellas begun in 2022:
  - a TV writer in a personal crisis
  - a lesbian literary editor being blackmailed
  - a church through the ages
  - different versions of a relationship
  - following a woman’s career at a corporation, from age 21 to retirement

These stories haunt me, every day. But I am waylaid with my work for Millennial Woes, so there is simply no time to get back into these stories and finish them. However, that is not the point. The point is that every one of them is very “visual”. The sci-fi romance, in particular, would be extremely adaptable to the screen. I have every aspect of it very clear in my head, pictured while I was writing it. I know exactly how every part of it “should” look. The novella about the church - inspired by an Erasure song - would be a beautiful piece if realised. Again, I have it clear in my head.

The question, of course, is how likely it is that these things will ever be brought into the real world. Leaving aside my own personal obstacles to writing them, adapting them is surely a pipe dream. What TV or film company would have any interest in my work? Even without the connection to me, would these stories have any appeal for the cretins of today’s media industry? It seems extremely unlikely. And, if by a miracle they did adapt one of them, I am certain they would wreck it; the priors of today would compel them.

But, with a sufficiently controllable AI imaging tool, I could make all of these myself, on my own. It would be a lot of work but it could be done.

---

Pondering AI in recent months, I have tried to imagine what obstacles, if any, would remain for drama-makers. In what ways will we still need “an industry” to produce, or aid in producing, these things?

The obvious area is actors. I have some experience of getting AI to render dialogue. It was for the female “train station announcer” in the Millenniyule 2023 trailer. For each line she said, I had to get many different renditions from the AI, then cut those up to use the most appropriate bits. It was onerous. The problem was that the AI doesn’t understand what is being said, so has no idea of what intonation to use for any given line of dialogue, and how to vary the intonation according to the character’s mood, situation, etc. Perhaps some kind of notation system could be developed, similar to what we have for music, that would instruct an AI how to “render” dialogue, with moments and sections of a certain tone being markable in tandem with the text - that tone then being tent-poled between key words. Perhaps this “notation” could be derived automatically by software from a recording of one’s own voice, then finetuned and applied to the character. Or, possibly, we will still use real actors. I imagine a mix of the two being standard practice. A real actor’s performance could be augmented after the fact rather than tediously (and expensively) bringing them back to the studio for re-takes; this would also be how a simple word was changed to accommodate a script revision. Meantime, sole producers would rely almost completely on AI.

(Even while I write this essay, technological development has advanced towards this sort of work. A Chinese company called ByteDance have unveiled software called DreamActor-M1 which enables one actor’s performance to be transposed onto a different actor. One interesting benefit to this is that it eliminates the danger for actors in making something politically dangerous, since, even though their original performance was real, they would not be recognised as the performer and thus their career, even though it wouldn’t be helped, would also not be destroyed. Immediately, this creates new possibilities.)

Another area where humans will still be desired is soundtrack composition. I would not want to instruct software to generate “a grand orchestral soundtrack in the mid-Romantic style” and then tolerate what it came up with. Music seems like something that has to be - that should be - conceived by a person, no matter how well software can do it. For the same reason I hate the idea of the script being dreamt up by software, music composed by software just seems utterly wrong. The tool must never become the author. The tool must always be completely subservient in the creative realm, used by the author to realise his intent. The intent must be his, never the tool’s.

For the rest - set design, costume design, director of photography, sound effects - it would really be a case of “how much do you want to do yourself?” It would be a choice, not a limit. You could do it all yourself. The deciding factors would be how much of your life you were prepared to spend on this one project, and how good you considered yourself to be in the various areas involved. Very few people are that multi-skilled and persistent, so I suspect a lot of specialists will still be gainfully employed even in an era of AI imaging - but they will be employed by a lone drama-maker to take some weight off his shoulders, not by a huge production company.

I can imagine AI stepping in again, to replace some or all of these specialists depending on the drama-maker’s resources. In the same way that a film director will ask a costume designer to come up with some ideas, the drama-maker will not want to do months of research himself for this one aspect of the drama, so will instead ask the AI to make some suggestions, which he will then choose from and tweak, etc. The process will be very similar, except that the human element has been replaced by AI. But did we ever care about the will of a costume designer? Or did we care that the director got what he wanted?

A related point is that it can actually be dangerous to get too involved with the minutiae of a project. Doing so frequently distorts the goal. The more involved the work is, the more distorted the goal becomes, and the more mental fortitude is required to prevent this happening. This is the tug-of-war between the macro and the distracting, niggling micro. This occurs not just with minor aspects of a production, but even in the act of creating some large aspect - that work can be so involved as to make one forget what one was aiming for in the first place. Sometimes this can be serendipitous - one ends up stumbling on some possibility that one hadn’t foreseen - but other times it can be demoralising and tiring. Again, AI could eliminate this problem by taking a lot of “the donkey work” off the shoulders of the drama-maker.

But then there is the problem of distribution. What will actually become of the dramas that are made in this way, by some nobody with a computer? Instead of just being things he shows to his (half-interested) friends, how do we get these dramas to a large number of people, and ideally to become “part of the culture”? This has been a problem ever since Fruityloops enabled the penniless nerd to make music in his bedroom and webcams enabled the politics sperg to make videos in his garage. The problem is that there is a massive abundance of “content” and the viewer has no way to sort the wheat from the chaff. This is why YouTube’s famous algorithm was a gift to the world (before they crippled it with censorship): it enabled people to find the videos they were looking for, or didn’t even know they were looking for. Now, would being distributed in this way recoup the costs of making the drama? Actually yes, in many cases, because those costs will be tiny. In an ideal future, you would also have local cinemas screening films made by people in the area. This would be a way to revitalise local cultures and identities that, currently, are being dissolved by the homogenising effects of globalism. Imagine seeing ten films every year that are set in your city, or even your town, and which record your culture, show your accent, depict your landscape, and address the issues that only your community care about. Such a thing is now very conceivable.

This doesn’t eliminate the problem of the established (or “legacy”) production companies being able to pay for advertising and algorithm prominence. But, as ever, they will be targeting the masses and the masses will be seeking them out. This exploitative master-slave relationship is one that I don’t think can be disrupted. Short of a political dictatorship, we just have to accept that the masses will always consume “slop” - gratefully and un-self-consciously - and leave them to it, and likewise leave the purveyors of said slop to purvey it. Something better - a situation where the masses aspire to high culture - would require an actual civilisation, and we don’t really have one any more. What we do have, however, is a minority of people who want to produce new art and drama, and I do believe AI will make that a thousand times more possible for them than it has been hitherto.

In such a world, the skilled drama-maker will be seen as the man who can do with imagery what a great author can do with words - craft a scene, pace an event, illustrate a mood, hint at a dynamic, construct a setting, convey an environment. All of the responsibilities, which have hitherto been split across hundreds of individual craftsmen, will merge into the singular domain of the drama-maker just as they have always been the singular domain of the author. The drama-maker will be operating in that total sense, that wholistic sense, responsible for the grand result but also for every tiny detail, just like an author. And it will not be easy. Sure, anyone will be able to generate a drama, just as anyone can point a 4K camera at traffic and automatically get “a high-definition shot of traffic”. But to do it well, to produce something that is not merely technically proficient and technically beautiful but actually good, will still be the preserve of the artist, unmistakably talented and obsessively engaged.

In that sense, AI imaging will be like every other new technology that came before it. With every tool, there are crappy, soulless, lazy and Philistine ways to use it, and there are dignified ways. With every tool, there is the risk of it taking over and dictating how the job should be done and what the job’s end result should be. These are risks that have plagued us since the beginning of time. I do not mean to deny the unique problem of AI (and remember I am speaking exclusively about AI imaging, not chatbots) but to remind us all that it is our choice how we use this and it is our moral duty to make the correct choice.

---

Regarding the technology itself, in a way we are lucky: we don’t have to make any choices at all, because this is happening whether we like it or not. No matter what we decree, proclaim or believe, other people will use AI imaging and our lives will be saturated with the fruits of it, good and bad. To complain about it would be like complaining about the weather; to try to keep it out of your life, utterly futile. Probably you and I are already daily seeing things - adverts, posters, images - that were created with AI, and we don’t even know it. The transition is happening… has already happened.

You can, of course, resist AI imaging in the sense that you can choose not to use it yourself. There is moral virtue in that choice, just as there is in a craftsman insisting on making his own tools. But I increasingly believe that there will not be a lack of moral virtue in choosing to use AI imaging.

After all, you didn’t build the components in the device you are reading this on, and neither did I. In a way, we have been in the realm of “magic” for decades now, even centuries. But there will be future “magic”, and we can only guess at it and how it will affect our lives. I believe that we should not condemn each other for using it, but hold each other to a standard in how we use it.

It is a tool. No matter how useful it is, it must remain that. The tool can enable us to make new things. But we must always remain in charge, faithful not just to the project, and not just to our own creative ego, but also to the shimmering original idea which compelled us in the first place.