Deep Fake Culture
The implications of Harold Ramis as a ghost...
The Disney-era Star Wars franchise used computer generated imagery to bring Carrie Fisher back as Princess Leia and to keep Peter Cushing working as Grand Moff Tarkin. Ghostbusters: Afterlife uses CGI and archival footage to render Harold Ramis into the film as Egon Spengler. A documentary about chef Anthony Bourdain employed artificial intelligence to generate audio of Bourdain speaking works he had written but never uttered. As technology and techniques improve, we will inevitably move away from the uncanny valley and towards convincing performances by artificial actors.
Middlebrow finds these practices somewhat acceptable and somewhat disturbing. If people want to see a holographic Tupac Shakur perform at Coachella, it seems fine to give that to them. But art and culture should always be moving forward, so perhaps we should not allow our nostalgic tendencies to keep us from experiencing and enjoying new things. Everybody has their favorite cinematic James Bond but putting new actors in the lead role is how the franchise endured. It both makes new stars and updates an old character for changing times.
In college in the 1990s, Middlebrow had a favorite professor who pitied young novelists because, he said, Philip Roth, John Updike, Saul Bellow, and the like would just not get out of the way. But, it turned out there was room for a whole new generation – Jennifer Egan, David Foster Wallace (who famously took on Updike with a savage review that likened the master to “a penis with a thesaurus”), Michael Chabon, Dave Eggers, Jonathan Franzen and Jonathan Lethem, among others, all rose to prominence despite the durability of the legends.
Imagine a literary equivalent of CGI. What if, rather than having to wait for infinitely typing monkeys to hammer our Hamlet, somebody created a program that reliably churned out new stories in the style, manner, and skill of Shakespeare? If we can automate the function of the actor or musician, then we can eventually automate the composer, the film director, the prose stylist, or painter.
The world is full of people who think their crafts and professions cannot be re-assigned to computers and these people are generally wrong whether they’re chess players or stock analysts. While it’s tempting to say that at least humans will always have a role in programming all this simulated creativity, the promise of machine learning is that the carbon-based first mover will not always be necessary. Can a program learn to act like Carrie Fisher without ever picking up a copy of Uta Hagen’s Respect for Acting? Can programs build other programs?
For now, we’re in the realm of tributes, easter eggs and one-offs when it comes to simulated entertainment. But where is this headed? Middlebrow imagines that he might one day access a cloud-based program where he can cast and watch his own films, using whatever talents he’d like to see, even if my whims lean into absurdities like Bill Murray as Rocky Balboa. All entertainment will be like Super Mario Maker.
Which all gets to the heart of the lost Middlebrow cultural experience – since we’re no longer all reading the same books, watching the same shows, or listening to the same music, our experiences become highly personalized and Balkanized. While this serves our tastes, it separates us from one another and renders communication and understanding more difficult. This all seems like a heavy thing to lay at the feet of the sweet the digital ghost of Harold Ramis. But if we are entering a world of infinite creative resources, where no entertainer needs to stop working because they’ve died (or even be born to start working) then there are going to be consequences to how we tell, understand, and share stories that will have a profound impact on what it means to be human.