Reports of an AI drone that ‘killed’ its operator are pure fiction

This week, news flew around the world that an AI-controlled drone had “killed” his boss. In a story that feels like a ripoff from a sci-fi thriller, this highly ambitious AI is trained to destroy surface-to-air missiles only with the approval of a human overseer, which is denied. Triggered the handler.

No wonder this story sounds like fiction. Because, in fact, it is. The story emerged from a Royal Aeronautical Society report describing a presentation by United States Air Force (USAF) Colonel Tucker Hamilton at a recent conference. The report pointed out that the incident was just a simulation, that there were no real drones, and that there was no real danger to humans, a fact that many high-profile headlines missed.

It was later found that no simulation had even taken place, the US Air Force issued a denial, and the original report was updated to reveal that Hamilton had “misrepresented”. The apocalyptic scenario was just a hypothetical thought experiment.

“The Air Force Department has conducted no such AI drone simulations and remains committed to the ethical and responsible use of AI technology. The Colonel’s comments were taken out of context and intended to be anecdotal.” It seems,” said a U.S. Air Force spokesman. insider.The U.S. Air Force did not respond new scientistRequests for pre-publication interviews.

This story This is just the latest in a series of dramatic tales about AI that at one point borders on hysteria. March, time The journal published a commentary by researcher Elise Yudkowski, in which she said that the most likely result of building a superhumanly intelligent AI would “literally kill everyone on the planet.” ‘ said. Elon Musk said in April that AI has the potential to destroy civilization, but a recent letter from AI researchers warned that the threat of extinction is so high that dealing with AI is a pandemic. and nuclear war should be a priority.

Why do these stories get so much attention, and why do we want to believe them? are being disseminated by AI experts to give authority to AI,” but it is worth noting that not all AI researchers share this view.

Beth Schinler, of the University of Zurich in Switzerland, says the media has an obvious motive for making such claims: “Fear drives clicks and shares.” But humans also have an instinctive desire to tell and hear scary stories, she says. She said, “AI looks like science fiction at first glance, but it’s also a horror story you want to whisper around a campfire. Horror stories are thrilling and captivating.”

One obvious driver of these stories is the lack of understanding about AI. Despite the fact that many have used his ChatGPT to create Limerick and Midjourney and conjure up images, few know how the AI ​​works under the hood. plug. AI is a familiar concept that has been around for decades, but the reality is that today’s advanced models are demonstrating capabilities that amaze the public, as well as experts.

“AI is very opaque to the general public,” says Singleler. “Education widely about AI’s limitations may help, but our love of post-apocalyptic horror stories may still come through.”