The media still can’t get AI right, and it’s getting worse

I’ve been watching the media coverage of AI for years now, and I’m tired. Not of the technology itself, but of the way journalists keep recycling the same tired tropes. The Facebook bot panic from 2018 is a perfect example, and it’s worth revisiting because nothing has changed.

Back in June 2018, five researchers at Facebook’s AI Research unit published a paper showing how bots could simulate negotiation-like conversations. Mostly the bots chatted coherently, but occasionally they’d spit out nonsense like “Balls have zero to me to me to me to me to me to me to me to.” The researchers realized they’d forgotten to constrain the bots to proper English, so the agents developed a kind of machine patois. Interesting, sure. Groundbreaking? Not really.

Then Fast Company ran with “AI Is Inventing Language Humans Can’t Understand. Should We Stop It?” They framed the whole thing as if the bots were running rogue and the researchers had to yank the plug. The story went viral. Suddenly every content farm was running with “Facebook engineers panic, pull plug on AI after bots develop their own language.” The Sun even compared it to The Terminator.

Zachary Lipton, a machine learning professor at Carnegie Mellon, watched this unfold and called it the “AI misinformation epidemic.” He wasn’t wrong. The research was interesting-ish, and the press turned it into sensationalized crap.

This isn’t new. Go back to February 1946, when the Eniac was unveiled. Journalists called it an “electronic brain,” a “mathematical Frankenstein,” a “wizard.” Physicist DR Hartree tried to calm things down with a sober explanation in Nature. The London Times ignored him and ran with “An Electronic Brain: Solving Abstruse Problems; Valves with a Memory.” Hartree wrote a letter to the editor insisting the machine was “no substitute for human thought.” Didn’t matter. It was the brain machine forever.

In 1958, Frank Rosenblatt showed off the perceptron, a rudimentary machine-learning algorithm that could recognize limited patterns. The New York Times declared it an “electronic brain” that could “teach itself” and would soon “be able to walk, talk, see, write, reproduce itself and be conscious of its own existence.” That’s a lot of hype for something that barely worked.

The pattern is depressingly consistent. Hype generates funding, researchers get paid, then reality sets in. By the end of the 1960s, pioneers like Marvin Minsky realized they’d underestimated the problem. Minsky had said machines would surpass humans in his lifetime. He co-authored a book proving Rosenblatt’s perceptron had fundamental limits. The hype cycle collapsed into the first AI winter.

Fast forward to 2018, and we’re doing the same dance. Social media has given us self-proclaimed “AI influencers” who do nothing but paraphrase Elon Musk. They cash in on the hype with low-quality pieces that get retweeted into oblivion. The result is dangerous. It creates unrealistic expectations, which leads to disillusionment, which threatens funding for actual research.

I’ve seen this cycle repeat enough times to be cynical. The media loves a good Frankenstein story. They love the idea of machines running wild, developing their own language, becoming self-aware. It sells clicks. But it does real damage to public understanding of what AI actually is and can do.

The truth is, AI is a tool. A powerful one, sure, but it’s not magic. It’s not about to wake up and start a war. The bots that said “balls have zero to me” weren’t inventing a new language. They were just broken. And the researchers didn’t panic. They fixed a bug.

But that doesn’t make a good headline, does it?