AI Has Become One Big Joke

People are already tired of machine-generated text, and they’re not afraid to say it.

If you want to really hurt someone's feelings in the year 2023, just call them an AI.

An all-star cast of celebrities and public figures have recently been the victim of such jokes: the NBA player Jordan Poole ("AI Steph Curry"), Raquel Leviss from the reality-TV show Vanderpump Rules ("what would happen if you asked chat GBT [sic] to create an American girl"), Transportation Secretary Pete Buttigieg ("our first A.I. cabinet member?"). That these slights span the three pillars of American life—sports, politics, Bravo—suggests that no one, or rather nothing, is safe. Such digs have popped up all over social media; on Twitter alone, insults like these have been levied against TV shows, songs, sports uniforms, commencement speeches, White House press releases, proposed legislation, and lots of news articles.

That AI has become an attack is a result of the huge moment for AI we’re in. Anyone can ask ChatGPT to write a silly poem or a college paper or a wedding toast—and that the chatbot can actually mimic human language with impressive precision is exactly why bots have taken off. But compared with human-generated work, much of what a chatbot spits out is dull or uninspired, riddled with clich​​és and recycled ideas. At a time when AI is capable of more than ever, Did a chatbot write this? is not a compliment. It's a diss.

"Are we sure that ChatGPT didn't write season three of Ted Lasso?" Bloomberg's Ashlee Vance tweeted. The Guardian, reviewing the action film Ghosted, called it "so carelessly and lifelessly cobbled together that we’re inclined to believe it's the first film created entirely by AI." My colleague Spencer Kornhaber, writing about Ed Sheeran's new album, surmised that its dull lyrics could’ve been composed by "a neural network trained on Sheeran's past work."

I called up Frank Lantz, the director of New York University's Game Center and a writer who has focused on making sense of AI, to discuss this style of joke. "Well, first of all, it is a funny insult," he said. Lantz positioned the dig as part of the larger arc of AI in pop culture. Even before ChatGPT arrived, robots were usually portrayed as both smart and dumb—sentient calculators that struggle to understand human emotion. He made beeping noises and put on a robot voice: "It does not compute."

Today we’re teasing ChatGPT and other bots for a different reason: because they sound like they’re posting on LinkedIn all the time, which is a different kind of soullessness. What has changed is that bots no longer primarily live on servers in some distant researcher's lab; they’re right in front of us, just one browser window away. "For a long time, people may have had concerns about [AI technology], but there wasn't a popular narrative that it was janky or broken or a thin version of the original," Meredith Whittaker, the president of the messaging app Signal and the chief adviser to the think tank AI Now Institute, told me. "And I think that's in large part because there wasn't a direct interface that most people in the public could play with and actually experience."

Part of why AI is an insult is also that we’re in the middle of an AI hype cycle, where every company is trying to stuff anything and everything into a chatbot. These jokes serve to ground us in the technology's present-day abilities: A bot can write a college paper, but can it write a good college paper? Janelle Shane, who runs the blog AI Weirdness, which tracks machine-learning quirks and gaffes, sees humor as a leveling force. "In some ways, this being a meme is a hopeful sign," she told me. Shane thinks it is somewhat useful to "bring that discourse back to the level of ‘No, no. These things are incompetent.’"

Read: Welcome to the golden age of clichés

But there's a darker spin. Humor, of course, is a coping mechanism; jokes about AI are on some level an expression of the anxiety around these tools. Bots are already replacing some jobs, and surely will replace more: Just last week, the National Eating Disorders Association announced that it was firing the humans who run its hotline and using a chatbot. The labor issues surrounding AI are also a big tension in the ongoing Hollywood writers’ strike, even as the public jokes about lackluster scripts being the work of AI. One picketer held up a poster that said A.I. THIS SIGN WROTE. Even a bad version of AI can take jobs, Whittaker argued, "not because it's competent, but because it will allow companies to justify degrading their position, paying them less, offering fewer benefits, turning them into contractors—all of this."

Over time, perhaps AI will get the last laugh. AI Steph Curry will cease to be a punch line if AI work becomes indistinguishable from the work of real humans. And the insult is becoming familiar enough that it verges on botlike itself. Last week, Fletcher Peters, an entertainment reporter for The Daily Beast, tweeted that writers need to find a better criticism—because "your review sounds like it was written by chatGPT."