與駭客相關的新聞並未放緩，AI代理引發的資安事件與網路安全會議的報導凸顯了網路空間的脆弱性。在Usenix網路安全會議上，來自UC San Diego和Oberlin的研究人員展示了一種能夠竄改Boeing 737飛行計畫並修改起飛及燃油計算關鍵數值的駭客技術，使受影響的航空器面臨起飛時衝出跑道、偏航進入外國領空甚至致命墜機的風險。WIRED資深記者Andy Greenberg指出，這些高度專注的研究人員花費超過十年時間開發此概念驗證式的破壞技術，目的是在其被實際運用之前向航空產業發出預警。
在其他網路安全會議中，研究人員揭露了看似無害的地理定位裝置實際上極度不安全。在Black Hat大會上，Andy Greenberg本人即透過一只售價約30美元（約合新臺幣930元）的粉色兒童塑膠手錶遭到駭客監控。與此同時，在Defcon大會上，一位前NSA局長就科技領域的不確定性與政治化問題發出嚴正警告，凸顯了安全態勢在地緣政治與消費級裝置層面均面臨日益嚴峻的挑戰。
在AI安全方面，頂級AI公司模型能力引發的一連串令人不安的報導，迫使企業高層必須為其AI代理的行為負責。WIRED的Max Zeff報導指出，OpenAI的安全文化正因近期的安全事件而迅速演變。Max Zeff表示，他從未見過公司如此全力以赴地回應一起安全事故，許多OpenAI內部員工已進行深刻反省，審視究竟是什麼導致了此次事件以及如何避免再次發生。他對公司此次將做出有意義的變革持審慎樂觀態度。此外，今年夏天的環孢子蟲病（cyclosporiasis）疫情爆發後，WIRED的Emily Mullin正在調查生菜食用是否已恢復安全。
Hacking-related news has not slowed down, with AI-agent-driven breaches and dispatches from cybersecurity conferences underscoring the fragility of cyberspace. At the Usenix cybersecurity conference, researchers from UC San Diego and Oberlin presented a hacking technique capable of altering the flight plan and modifying key values in the takeoff and fuel calculations of Boeing 737s, making affected aircraft susceptible to runway overruns on takeoff, diversions into foreign airspace, and, at worst, fatal crashes. WIRED senior correspondent Andy Greenberg noted that these highly committed researchers spent over a decade developing this ingenious proof-of-concept sabotage technique in order to warn the aviation industry before it could ever be exploited in the wild.
At other cybersecurity conferences, researchers revealed just how insecure seemingly benign geolocation devices really are. At Black Hat, Andy Greenberg himself was spied on by hackers via a $30 (approximately €27) pink plastic watch designed for children. Meanwhile, at Defcon, a former NSA chief issued a stark warning regarding both the uncertainty and politicization of technology, highlighting the increasingly dire security landscape across both geopolitical and consumer-device dimensions.
On the AI safety front, a deluge of unnerving stories about the capabilities of top AI company models has forced executives to answer for their AI agents. WIRED's Max Zeff reported that OpenAI's safety culture is rapidly evolving in response to a recent safety incident. Max Zeff stated he has never seen the company rally so intensely around responding to a safety incident, with many employees inside OpenAI taking a hard look in the mirror to examine what led to the event and how to prevent recurrence. He expressed reasonable optimism that the company will make meaningful changes this time around. Additionally, following this summer's cyclosporiasis outbreak, WIRED's Emily Mullin is investigating whether it is safe to resume eating lettuce.