Why Screenwriters Rebelled Against the 'Dystopian Future' of AI-Evaluated Scripts
"You can't quantify a Daniel Day-Lewis performance when you enter a script into this software."
Image: Steven Tsapelas
On April 18, screenwriting website The Black List announced its newest tech initiative: an AI which would "read" and evaluate users' screenplays. The Black List, by its own admission, expected some skepticism.
But it couldn't have possibly prepared for the sheer volume of negative criticism. The backlash was so overwhelming that, only a day later on April 19, the website took down the option to use the tech. Founder Franklin Leonard also posted a note on the website, which read in part:
"We were, are, and will remain an organization that celebrates, reveres, and often stands in awe of writers, and today is no different. No matter how difficult it was to hear some of the feedback about our decision (I will say that 'You'd fit in well in the Trump administration' felt unnecessarily hurtful and inaccurate,) we sincerely appreciate it—all of it—if for no other reason than that it allows us to better serve."
The Black List started, as its namesake implies, as a list. In 2005, Leonard surveyed 100 film industry development executives for their favorite feature film scripts that had yet to be produced. In the intervening years, the annual list has become an industry distinction; award-winning scripts such as Slumdog Millionaire, Argo, and Spotlight all appeared on the list before becoming critically acclaimed films.
The Black List has evolved into a community of screenwriters and filmmakers; it organizes local meetups. It runs writing workshops and maintains a library of resources on the craft of screenwriting. And most notably, it offers members screenplay evaluations by professional readers, who will rate the scripts and give them constructive feedback. Those screenplays are also uploaded online, where producers and filmmakers can filter and search for them by type.
The now defunct script analysis tool was rolled out in partnership with technology company ScriptBook. Via machine learning and natural language processing, the AI "learned" thousands of produced screenplays. It then used the resulting algorithm to evaluate—objectively, allegedly—screenplays that were fed into it. Script analysis of this sort, according to The Black List, is not new:
"Increasingly, these tools are being used by studios and production companies to make decisions, so we want to offer such a tool to writers at the lowest price point possible."
The 4-page summary report that users would have received covered several practical areas, such as a film's potential genre, rating, and film budget:
It could also measure the protagonists' and antagonists' likeability:
And it could also do Orwellian deep dives into each characters' "sentiment," like this breakdown of three characters' anger, annoyance, fear, and happiness.
The price for one of these reports? One hundred dollars, which, interestingly enough, is a higher price point than a Black List professional reading, which costs $50 (plus the cost of membership, which is $25 per month).
I feels like this is a dystopian future where a computer is judging the merits of a screenplay.
"What's disheartening about a [tech] program like this is that when you're trying to break into the industry, everything already feels like a racket already," said writer and filmmaker Joyce Wu in an interview with Motherboard. "There's gatekeepers purporting to give you a stamp of approval, and it's preying on people. As a screenwriter, I feels like this is a dystopian future where a computer is judging the merits of a screenplay. It's hard to not have a knee jerk reaction to that."
"The vast majority of people who would be paying for that service would not need the information it would have provided," continued Wu. "[The studios] do that stuff on their own anyway."
Wu is an independent filmmaker who wrote, directed, and produced her first feature film, She Lights Up Well, in 2014. Since then, she's been writing, directing, and starring in a web series entitled Mr. Right, which she's currently shopping as a pilot and hopes to convert into a television series. As an Asian American woman in Hollywood, Wu knows firsthand about the politics that determine what gets made and what does not.
"People want to believe that Hollywood is a meritocracy, and I wish that was the case," said Wu. "When a screenwriter doesn't get noticed, he or she might think, 'Maybe my script isn't any good.' And that insecurity feeds into the 'pay-to-play notion'—that maybe if I paid people to analyze my script, I'd get better feedback."
"If the writer does not have a high quality script, the information that this analysis supplies would be useless. As you have said, the analysis makes no judgement on the quality of the script, just the relation of its elements to other films.
So prospective customers should really be informed of this and that they should not pay for this analysis unless their script has already been rated highly on the [Black List] website. Otherwise, they are paying for information that is useless to them as they don't have a script that could be sold, regardless of what this analysis tells them."
Other users, such as coquinbuddha, felt that such a tool promoted a wrong-headed mentality about script writing—there is no quick fix to becoming a successful writer. One must do the necessary work and research.
"This may sound elitist, but if a person can't even figure out what genre their script is, they aren't very likely to have a career in this industry.
With respect, if you really want to help new writers, don't encourage them to believe there is some sort of alternative to learning the craft and business behind being a professional writer. Being a professional writer requires research, diligence, actually learning how things work. Telling people who haven't bothered to understand how to actually do the work of a screenwriter -- but have big dreams of making it someday -- that now there's this report that can tell them everything they were too lazy to learn, isn't really helping them… there are no $100 shortcuts."
Steven Tsapelas is one of those writers with big dreams. He currently works as a supervising writer and producer for MoPo Productions, but he has also tried, with varying degrees of success, to break into mainstream entertainment. His web series, We Need Girlfriends, went viral in 2006, around the same time that YouTube first took off. It eventually became a TV sitcom pilot for Sony Pictures, but never got picked up. Tsapelas recently submitted a new screenplay, Ugh Whatev Okay, to the Austin Film Festival. He describes it as a homage to 90's teen comedies, with a twist.
Even at his stage of the game, where he's on the verge of breaking out, Tsapelas is against using an AI program like this one. He'd rather seek out professional feedback for the type of money he would pay.
You can't quantify a Daniel Day-Lewis performance when you enter a script into this software.
"There are rules that a lot of amateur writers don't follow when they write screenplays," said Tsapelas in an interview with Motherboard. "And it is important to have some of those guidelines in place; if you submit a screenplay [as an emerging writer] that's going to cost $500 million, you're going to look like an idiot."
"But a lot of this is also a guessing game," continued Tsapelas. "Nobody knows what's going to hit or work. And when you talk about the likeability of the character, how do you quantify that? There Will Be Blood has the least likable character of all time. But you love him, and you can't quantify a Daniel Day-Lewis performance when you enter a script into this software."
Tsapelas also echoes previous sentiments—that writers who take their craft seriously will learn fundamental information the hard way, through practice and experience.
"If you're a working writer, you know what you're writing," said Tsapelas. "You know how to craft your pitch."