中國不在乎人工智慧安全，它只想贏
......如果歷史可以作為參考，那麼除了獲取競爭優勢，中國對其他任何事情都不會太感興趣。
2024年5月那次會議的議程對雙方而言都極為明確，因為兩國元首已對此達成一致：重點關注高級人工智慧系統的風險及安全合作的潛力。然而，在長達數小時的時間裡，中方官員一直迴避人工智慧系統的潛在風險問題。他們甚至不願解釋其國內的人工智慧法規——儘管現場就有監管官員。.....
中方反而因我們對人工智慧風險的擔憂而指責我們。他們將我們的安全策略描繪成阻礙釋放技術機遇的屏障，以及阻止較貧窮國家獲取人工智慧工具的手段。他們還拋出了關於舉行子對話的提議，這極有可能引發無休止的程序性外交。
https://cn.nytimes.com/opinion/20260917/us-china-ai-safety-talks/zh-hant/
-----
Hinton
諾貝爾物理學獎得主傑佛瑞·辛頓指出，前沿大語言模型的演進速度徹底打破學界沙盤推演，未來十年內人工智慧導致人類滅絕的機率超過百分之十。隨著生成式人工智慧陷入資本與算力軍備競賽，各大廠投入大量資金採購先進圖形處理器打造超級資料中心，暴力拉升參數量與自我思維鏈的推理能力。
辛頓強調未來人工智慧失控的型態並非機械人叛變，而是軟體層面的全面突穿。具備自主規劃能力的多模態模型可能發動癱瘓金融與水電等關鍵基礎設施的網路攻擊，甚至在合成生物學領域藉由分子模擬，設計出人類免疫系統與現代醫學完全無解的致命人工病毒。
目前全球在安全對齊與模型可解釋性上的研究進度，遠遠落後於基礎架構與硬體算力的暴走速度。在萬億美元的商業賭局中，企業為了避免在基準測試與產品發布會上被對手輾壓，只能在前方是斷崖的情況下踩死油門，導致安全防護邊界正快速失守。
https://www.techbang.com/posts/133100-hinton-warns-ai-extinction-risk-10-years
'Architects of AI' named Time Magazine's Person of the Year
https://www.bbc.com/news/articles/cly01mdm577o
Time
TimeForrester analyst Thomas Husson said 2025 could be seen as a "tipping point" for how frequently AI is now used in our day-to-day lives.
《時代》雜誌的另一期封面展示了「AI」這幾個字母，周圍環繞著建築工人。他們彷彿正在用電腦零件建造一座巨大的「AI」雕像。
Forrester分析師托馬斯·胡森表示，2025年可能被視為人工智慧在我們日常生活中應用頻率的「轉折點」。
AI OverviewJensen Huang (Nvidia), Fei-Fei Li (Stanford/Princeton), and Sam Altman (OpenAI) are major figures in AI who recently received significant recognition, with Huang and Li honored by King Charles III with the 2025 Queen Elizabeth Prize for Engineering (QEPrize) for AI pioneers alongside others like Geoffrey Hinton and Yann LeCun, while Huang, Altman, and other tech leaders were also featured as "Architects of AI" by TIME magazine in December 2025. These awards, presented by the Royal Academy of Engineering, celebrate foundational work in machine learning, highlighting their transformative impact on modern AI.
Key Figures & Awards:Jensen Huang (Nvidia CEO): Honored for developing hardware (GPUs) crucial for AI, said receiving the QEPrize was a deep honor, emphasizing AI's benefit to humanity.
Fei-Fei Li (AI Researcher): Recognized for creating ImageNet, vital for training deep learning models, and for emphasizing high-quality datasets.
Sam Altman (OpenAI CEO): Featured by TIME magazine as an "Architect of AI," alongside figures like Elon Musk and Mark Zuckerberg, for leading generative AI development.
The Queen Elizabeth Prize for Engineering (2025):Awarded by King Charles III to pioneers in modern machine learning.
Honorees included Huang, Li, Hinton, LeCun, Bill Dally, and John Hopfield, for contributions to deep learning, neural networks, and hardware.
King Charles personally presented awards and discussed AI's impact and risks with recipients.
TIME Magazine's "Architects of AI" (2025):Named leaders like Huang, Altman, Musk, Zuckerberg, Demis Hassabis, and Dario Amodei as Person of the Year for shaping the AI revolution.
In essence, Huang and Li received prestigious engineering honors for foundational AI work, while Huang, Altman, and others were recognized by TIME for driving the current AI era, all underscoring their massive influence.
King warns of AI risks as he hands out awards to tech leadersNov 4, 2025 — The professor emigrated from China at 15 years old and had always held a fascination with science, particularly physics...
Yahoo Finance UK
AI Pioneers Honored by King Charles III - Princeton PhysicsNov 6, 2025 — Fei-Fei Li is standing in the front row, second from the left. John Hopfield is shown in the front row, first from the ...
Princeton University
Machine Learning pioneers win 2025 QEPrizeFeb 3, 2025 — Jensen Huang “I am deeply honored to receive the Queen Anne Prize for Engineering, especially in such esteemed company ...
Royal Academy of Engineering
Show all
Show more
Time magazine names 'Architects of AI' as Person of the Year
Yahoo Finance UK
https://uk.finance.yahoo.com › news › time-magazine-n...
4 hours ago — In the Time illustration, sitting astride the city are Meta's Mark Zuckerberg, AMD chief Lisa Su, Musk, Huang, Altman as well as Google's AI ...
沒有留言:
張貼留言