Writing Is Not a Widget
Big Tech and Big Media want to sell you a lesser product.
At BuzzFeed’s Investor Day, CEO Jonah Peretti (fresh off shuttering BuzzFeed News and taking responsibility for its failure while still suffering no penalties for that failure) was upbeat about how his media company will harness AI to generate content:
Peretti’s comments about what AI can do are nonsense because it’s “magic box” talk. Big Tech and Big Media are excited about AI not because of its creative potential, but because of its labor potential. It’s Peretti’s stage at an Investor Day (a day to tell people while they should keep investing in his company, not be challenged on the specifics of those investments), so there was no challenge about what content that’s “dynamic with embedded intelligence” means or what exactly is a format that’s “more gamified, more personalized, and more interactive.”
What Peretii is really talking about comes two paragraphs later when he talks about making his creative teams, “more efficient and sustainably expands our output without increasing fixed costs.” To put it another way, when the AI cranks out article after article, it doesn’t pester you for a living wage or health insurance. It’s a machine. You may have to pay some part-time employee to glance over the copy to make sure the AI didn’t write anything egregious, but it doesn’t really matter because it’s simply about pulling a crank and out comes an article.
The investor class and C-suite executives aren’t excited about AI because of creative potential. They’re excited because it can make the line go up. People are messy. They make demands, they have free will, and they can go to a competitor if you don’t pay them enough. The nerve! But good old trusty AI wants for nothing. You feed it a prompt, and out comes an answer that looks like an article in that it has a semblance of syntax devoid of messy things like a personality (I assume a personality module will come later when you can get the AI to answer in the form of Mr. T or Hannibal Lecter).
If all you care about is what comes out at the end, then you don’t mind that AI is delivering an inferior product. And some websites have already wised up to use AI to game Google by overloading it with SEO-friendly articles written by a machine that will hit the proper keywords. These websites may be raking in traffic (and that’s clearly what Peretti has his eye on), but the quality is lacking, and at some point, if you offer me a choice between bland AI results and something written by a person, the person wins out because writing is a valuable skill.
A machine may be able to do something simple in a pinch (“What’s the tallest mountain in the world?”), but a writer can tell you what it’s like to climb that mountain. AI can scrape together other people’s answers, but it doesn’t understand what’s compelling in the information. This is like Major League Baseball deciding that since people love home runs, they should get rid of pitchers, put the ball on a tee, give mediocre players an aluminum bat, and see what happens. And what happens is you haven’t changed the game; you’ve ruined it.
The gamble here isn’t that AI can crank out articles or scripts or whatever. The question is whether the audience will accept it. While the proverb goes, “No one ever went broke underestimating the intelligence of the American public,” we’ve seen these supposedly tech-savvy ideas repeatedly blow up in companies’ faces. Remember pivot-to-video, which was a scam perpetrated by Facebook? Remember less than two years ago when you needed to get into cryptocurrencies because they were the next big thing? I think AI will have its uses, but it’s telling that Peretti and his ilk have immediately leapt to, “We found a way to cheapen labor costs and hope that the output isn’t so bad that it will turn off the audience.” Regardless of what you may think people are silly enough to buy, you actually can go broke underestimating the American public. The problem is that guys like Peretti never go broke; it’s the people who work for him that suffer the consequences of bad decisions.
So let’s try something else this time and not buy into the hype. Let’s not go the AWESOM-O1 route, and instead have industries where we value writing, value the reader, and guys like Peretti don’t have to pretend that they’ve invented a shiny new future for the investor class.
In this episode of South Park, Cartman dresses up as a robot to prank his classmate Butters only to have to keep the ruse going when he discovers Butters has incriminating footage of Cartman. This leads them to Hollywood where executives think that AWESOM-O is a real robot and can spit out ideas.