AI Might Be Your Friend After All
It’s that time of year again!
Actually, I don’t know why I said that. It’s always that time of year for something.
The good news is that, after a long, cold, hard winter (or two . . . or three), we’re finally seeing more work – production is showing signs of coming back after last year’s SAG-AFTRA and WGA strikes, followed by this year’s near-strike by IATSE. It takes a while to get things through the pipeline, but we’re seeing some early hints of it – like the buds of spring.
It will take some time – and perhaps longer here in New England – for production to return to the scope and scale we saw pre-COVID. But the pandemic had already done some real damage to our community, which we were still recovering from when those first strikes hit.
Those strikes were a necessary step – for all of us – as any realization of all the ramifications of Artificial Intelligence has just barely begun. There are elements to this that we simply can’t foresee, and there are some we can – all of them need to be addressed so those contracts can be future-proofed (to the extent possible) in anticipation of what AI might do.
There is the obvious stuff that we see now – actors being replaced or regenerated from random electrons or being made to do things they could never do. Or would never do . . . from deepfaking politicians to revenge porn to bringing the dead back to life, almost everything is on the table, and all of it is dark; completely unacceptable without the individual’s consent or without the community of actors standing up to say ‘no’ to artificially generated actors taking their jobs.
It was a similar issue presented by the Writers Guild – to keep their integrity as creators, it’s vital to stand against AI-fabricated ideas – or maybe entire treatments and scripts. AI doesn’t produce a script like a human – not yet, anyway – but that writing is on the wall (pun only slightly intentional). As with everything else AI, it’s simply a matter of time. Five years ago, even the idea that AI would threaten any of our jobs was inconceivable. No one really knows where we’ll end up in another five years, so it’s best to write in those protections now.
Below-the-line employees, represented primarily by the International Alliance of Theatrical Stage Employees (IATSE) and the Teamsters, had a tougher course of negotiating – AI hasn’t come for as many of our jobs (yet), but we need those protections, too. We already see it in the scaling back of practical effects work on set; as one example, there’s no need to actually blow stuff up or set it on fire when it can be convincingly (and more safely) done after the fact. AI could also cause the scaling-back of some practical jobs on set – lighting can be adjusted in post, maybe wardrobe gets adjusted (or replaced) after the fact, makeup can be at least tweaked, sometimes created out of whole cloth – and on and on.
It’s coming for sound as well; we’ve already seen what plug-in processors can do for audio enhancement and repair. It might be only a matter of time before we can set up a good-enough camera in an undecorated space with a satisfactory microphone and a skeleton crew of technicians to capture the moment for AI to enhance later.
I am not “anti-AI” – for example, I’ve used versions of automatic mixing algorithms that are a tremendous help, particularly with live shots. The software can open up a mic that needs to go live faster than I ever could, for example. It’s not a very big leap to imagine a time when we’re no longer dealing with a human mixer/operator, but telling a computer how we’d like that piece mixed. Throw in a little sound design while you’re at it, Siri – you know what I like.
AI, as it exists now, largely models its responses and generative solutions on pre-existing works – it’s copying everything from a previous human creator or a human body of work, from concept to style to execution. For AI to do its job, some human, somewhere, needs to have done something like that already; AI is simply cribbing from our work and creating its own amalgamation of that in its version of new, supposedly “original” work. In its current incarnation, AI technology simply emulates the human experience, and its learning cycle never end. More and more frequently, it’s good enough to fake us out – and those instances are going to occur more often as we progress down this path.
I, for one, look forward to our robot overlords taking the reins – just a little, maybe. My hope is that it will take on the more mundane aspects of what we do, giving us more time to concentrate on the serious business of creating. It can also help with new concepts, ideas, and combinations of elements, while finding entirely new approaches – an AI prompt or suggestion might trigger the beginnings of an idea, but any genuinely new creative territory (for now, anyway) will remain in the domain of the human creator.
It’s a new season . . .