Meta客服AI遭劫持敲警钟：模型防线之外的安全考题

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. The Meta hack shows there’s more to AI security than Mythos On Monday, reports emerged that attackers had used Meta’s AI customer support agent to steal Instagram accounts. Their approach was simple: they asked the agent to link the accounts to email addresses they controlled, and it complied. Since Anthropic announced that its Mythos model was too good at hacking for a general release, cybersecurity concerns have focused on the risk of superpowered AI systems overwhelming computer infrastructure. But the Instagram hack shows that far simpler exploits can still cause damage. As companies offload more work to AI, these comparatively unsophisticated attacks are becoming harder to ignore. Read the full story to understand why. —Grace Huckins Are AI chatbots making us lose control of our brains? Gloria Mark, a psychologist at the University of California, Irvine, fears that digital technologies are weakening our cognitive abilities. Her research suggests attention spans have fallen sharply over time, leading to higher stress and lower performance. She now believes AI tools like ChatGPT and Claude may accelerate this shift. “You’re deferring your cognitive work to AI,” she said. “And it’s not good for us.” Mark argues this could weaken critical thinking and emotional intelligence. Luckily, she thinks we can course-correct by changing our relationship with these technologies. Find out how AI could reshape attention and thinking. —Jessica Hamzelou This story is from The Checkup, our weekly newsletter giving you the inside track on all things biotech. Sign up to receive it in your inbox every Thursday. The must-reads I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology. 1 Anthropic has called for a global slowdown in AI developmentIt flagged the risk of models “self-improving.” (WSJ $)+ And wants a coordinated plan to stop them. (Reuters $)+ Skeptics note that the timing is awfully convenient. (The Register) 2 In a first, scientists have precisely edited human embryo genesThey relied on a newer gene-editing technique. (NYT $)+ Genetically-modified babies could be on their way. (Guardian)+ Companies have big plans for the technology. (MIT Technology Review) 3 US officials have discussed taking financial stakes in the AI firmsThey’ve held talks about the government acquiring shares. (Reuters $)+ Sam Altman pitched the idea to the White House last year. (WSJ $) 4 Bot web traffic has overtaken human web trafficCloudflare said 57.4% of traffic now comes from bots. (NBC News)+ Its CEO expected the milestone at the end of 2027. (CNET) 5 The White House plans to bring AI doctors into American medicineIt wants chatbots to diagnose illness and prescribe medicine. (WSJ $)+ But we don’t even know if healthcare AI actually helps patients. (MIT Technology Review) 6 Meta quietly added facial recognition code for smart glasses to its appThe exploratory feature would identify people via biometric data. (Wired $)+ Smart glasses are also entering warfare. (MIT Technology Review) 7 South Korea’s labour minister wants tech firms to share AI profitsKim Young wants staff and suppliers to get a share. (Reuters $)+ He helped avert a huge strike over AI profit-sharing at Samsung. (NYT $) 8 Canada’s highly-anticipated AI strategy has launchedIt promises over $2 billion in funding and aims to create 250,000 jobs. (BBC)+ AI could strengthen democracy. (MIT Technology Review) 9 Investment in agricultural tech is boomingThat’s good news at a time when we’re facing unprecedented levels of food market volatility. (The Economist $) 10 Bumblebees can use tools to solve problems, new research showsNot just busy—they’re clever too! (Guardian)  Quote of the day “Welp, that happened faster than I predicted.”  —Matthew Prince, co-founder and CEO of Cloudflare, one of the largest internet hosting services, reacts on X to reports that bots have overtaken humans in driving web traffic. One More Thing CHRISTOPHER PAYNE Inside the machine that saved Moore’s Law In a Connecticut clean room, the Dutch company ASML is developing the world’s most advanced machine for extreme ultraviolet (EUV) lithography, a crucial process for manufacturing microchips. The system has become vital to Moore’s Law—the observation that the number of transistors on a chip roughly doubles every two years as components shrink, driving gains in performance and efficiency. “Without this machine, it’s gone,” says Wayne Lam, a director of research at CCS Insight. “You can’t really make any leading-edge processors without EUV.” Discover how ASML’s EUV technology saved Moore’s Law. —Clive Thompson We can still have nice things A place for comfort, fun, and distraction to brighten up your day. (Got any ideas? Drop me a line.) + Tech bosses love Tolkien. Here’s what the writer might think of them.+ Rare footage captures an underwater volcano erupting beneath the Pacific Ocean.+ Watch a tiny rescued cub grow into adulthood in this heartwarming tiger compilation.+ This medieval version of “Take On Me” is like stepping into a tavern of synth-pop bards.