The second episode of Electric Dreams delves into an interesting nugget: What happens to a world designed for pure automation when that world finally ends? A nuclear war leaves a small group of humans struggling to scavenge and survive, but also having to reckon with the oppression of a megacorporation known as Autofac, whose AI guidance ensures that the factories never shut down. Meaning they not only continue trying to deliver goods, but serve customers who don’t even need them. For the residents of one particular small town, this isn’t a small inconvenience: The factory is still polluting the land, cutting them off from any exit, and pushing them even further toward the brink of survival.
Enter a young woman named Zabriskie (Juno Temple, and points for the Antonioni reference), who has a cunning plan to draw the Autofac AI out to their village, reveal what it’s doing wrong, and convince it to change in the name of better “serving” its customers. The townspeople fret and argue over this, while Zabriskie tries to distract herself with her sweet librarian boyfriend who’s clearly in love with her, while she’s more distant. Why is that, we wonder? But it turns out that Zabriskie’s plan works: Days later, “Alice” (Janelle Monáe!) shows up in the form of a PR robot straight out of Metropolis. They talk cordially, but Alice will not relent. She will not alter from the Autofac programming, and she will not stop echoing the hallmarks of every horrible automated-customer-service experience you’ve ever had. So Zabriskie has to come up with another plan — but rather than get into it, I’m just going to cut to the chase.
This episode is a narrative magic trick.
Which means we are put off from understanding what’s really going on, and not meant to really understand Zabriskie’s behavior until the last moment when everything is suddenly revealed. Because, as it turns out, Zabriskie and her compatriots are not real! They’re robots just like Alice, built so that the automation production pipeline can have happy customers! But … double twist! Zabriskie knew this all along! She discovered that she was the CEO of Autofac in a prior life, and now she’s uploaded a code that can undo it all! She’s learned the value of life and love even though she isn’t real! Her virus shuts down Autofac and she returns to her boyfriend, finally hugging him! Everything makes sense! What a great trick!
But no matter how logically it unfolds, the real question of “Autofac” is simple: How moving is the magic trick when you look at the path it took to get there? Just like with episode one, we are pushed away from understanding our main character constantly. We don’t know why Zabriskie wants to hide her magazine. We don’t know why she keeps thinking about her dreams. It is all meant to draw our curiosity, but not our emotional investment. This would be less problematic if the story moved with any kind of purposeful zeal before the reveal, but instead we get constant narrative delay. For instance, when Zabriskie first finds the magazine, it then flashes back to her discovering the same magazine, but there’s literally no other information. It’s as purposeless a flashback as I’ve ever seen. It’s not the only delay tactic: It takes nearly 20 minutes for Alice to appear. A character will say, “I love you,” and then rather than deal with it in the scene, someone else interrupts then. These delays aren’t purposeful — this isn’t a Hamlet-like meditation on the nature of delay itself — they’re pure plot-blocking because the story has to constantly invent ways to fill 50 minutes before getting to the main reveal.
This is the central problem with how we approach reveal-based storytelling, especially with anthologies like Electric Dreams and Black Mirror. If you’re basically just teasing that one reveal for an entire episode, you’re pushing your audience’s emotions away instead of bring them into the drama of the journey. It reminds me of something that Joss Whedon reportedly used to say about Buffy the Vampire Slayer: “Play your cards early, it forces you to come up with more cards.” After all, this episode’s most interesting moment comes during a confrontation between Alice and Zabriskie, when we learn that Alice is more human than she’s letting on and will put self-preservation first, so she will go along with the plan. It’s a moment that creates a real new conflict and changes the scope of the narrative. But rather than deal with it, “Autofac” just shelves the entire discussion until an elevator ride 15 minutes later.
This is the problem: If you have only one card to play at the end, you have to decide what a story seems like it’s about before it becomes what it’s really about. And both have to feel meaningful. Again, I point to the best episodes of The Twilight Zone, where the initial story makes total sense and is dramatically compelling even before the twist.
Of course, that’s not the only thing in “Autofac” that matters. This episode is ostensibly about the dangers of automation. It’s about the dangers of never thinking about the possible downside of technology; it’s about a Silicon Valley logic where convenience is everything and invention is its own reward. Alice tells us, “The autofac doesn’t need to understand. It can make anything it has the blueprints for.” And so the ending, while clunky, still points to the idea that we’ve all become robots who will buy into automation because it’s convenient, and because we feel we have no other choice. Zabriskie’s decision is a (sadly undramatized) mea culpa, the struggle to undo that which she has created. But the reality is that it’s just another riff on apocalypse porn. The fantasy of the apocalypse isn’t that we are doomed, but that we’ll survive it. That we’ll get to have an “I’ll told you so” to society, whether it be the issue of global warming, nuclear war, or automation. While “Autofac” at least has the courage to take on complicity in that idea, the notion of that complicity brings me to the episode’s meta-narrative, which is the only real thing about it that will keep me up at night.
Make no mistake, this is an episode about the dangers of automation and drones. And it’s being put out by … a company that leads the world in automation and drones.
I get the artistic notions within this. If you’re the creators, you want to get one over on your parent company and tell the truth to their face. If you’re Amazon, you want to show that you can play ball, take criticism, and not censor. But to what end? Is Jeff Bezos really going to pull a Zabriskie and undo all this? There’s a larger problem at play here: Acknowledging a problem and doing nothing to fix it is essentially a useless action. For Amazon, it’s hard to see this as anything but performative responsibility. Really owning the message of “Autofac” would require change. It would require taking on the actual cost of putting people first. But why do that when you can just put out an episode of TV instead?
If that’s not the most cynical thing in the world, I don’t know what is.