This week’s cover
The stunning success of ChatGPT, a chatbot that can write essays, pass law exams and even make bad jokes, has divided the world. Optimists think artificial intelligence (AI) will help us design new drugs or generate power by nuclear fusion. Pessimists fear that superintelligent machines will go rogue and wipe out humanity. Our cover package this week weighs the evidence. We needed an image that suggested both peril and promise. One early ideaa child hiding behind a sofa because shes scared of an old-fashioned robotmade us smile. In one way the analogy works well: peoples understanding of AI really is rudimentary, like a childs. But from a grown-ups perspective the robot looks neither scary nor likely to change the world. We needed something with more menace. Option two was a bit too apocalyptic. It showed a museum exhibit of the last human skull, calling to mind the Terminator films, in which the AI that runs Americas defence systems becomes self-aware and tries to kill everyone. In one scene, the camera lingers on a human skull just before a robotic foot crushes it. Another idea displayed the brain as a puzzle to be solved, suggesting the struggle to grasp how intelligence works. However, it wasnt quite dramatic enough. The prize for solving a ball-and-maze brain teaser is mild satisfaction; the penalty for getting it wrong is a couple of wasted minutes. We also tried putting a photo of Charles Darwin on the wall, gazing down on a robot whose pose echoed the statue of The Thinker by Auguste Rodin. We liked this one. Human intelligence evolved by chance, through natural selection. The artificial kind is being built deliberately, but the process is still unpredictable. And we played with the idea of showing wiring in the shape of a brain, connected to a human. Is it feeding useful information to people, or being used to manipulate them with individually tailored misinformation? We worked this up into a cover image, slightly tweaked so that the human was struggling to control a big, floating brain. But in the end we went with a horns-versus-halo theme. At first we thought we needed a headline. But then we decided it was bolder simply to show an angelic A and a devilish I, and let our readers old-fashioned human intelligence do the rest. Leader: How to worry wisely about artificial intelligence Essay: How AI could change computing, culture and the course of history Science: Large, creative AI models will transform lives and labour markets Science: Large language models ability to generate text also lets them plan and reason Science: How generative models could go wrong China: Can Xi Jinping control AI without crushing it?