18 Nov 2025明報#7546
港大論文引AI虛構文獻 出版社依倫理委會指引調查
An academic paper involving Yip Siu-fai, associate dean of the Faculty of Social Sciences at the University of Hong Kong, was found early this month to have used AI and included fabricated sources. Publisher Springer Nature previously said it had received a report from the authors and that the incident did not affect “the integrity of the article’s core conclusions”. Less than a week later, Springer Nature published an “Editor’s Note” on its website last Friday, saying it was conducting an investigation in accordance with the guidelines of the Committee on Publication Ethics. Yip said yesterday that he had “done everything I need to do” and would patiently await the investigation’s findings.
港大社會科學院副院長葉兆輝有份參與的學術論文本月初被揭發使用AI並出現虛構資料來源，出版社Springer Nature早前稱已收到作者報告，事件不影響「文章核心結論的完整」。事隔不足一周，Springer Nature上周五於官網發表「編輯的話」（Editor's Note），稱正根據出版倫理委員會的指引展開調查。葉兆輝昨回覆稱自己「需要做嘅事已做晒」，將耐心等候調查結果。
6 Aug 2025光傳媒#7140
A.I.滲透宣傳戰 中國公司借《國安法》操控香港輿論
The rapid development of generative artificial intelligence has ushered the global information war into a new era. US scholars specialising in international security affairs said Chinese technology company GoLaxy had become a “leader” in this field. After the National Security Law was implemented in 2020, it suppressed opposing voices in Hong Kong; in 2024, it intervened in Taiwan’s elections, and it is now preparing to enter the US market, posing an unprecedented threat to democratic systems worldwide.
Brett J. Goldstein and Brett V. Benson, professors at Vanderbilt University, co-authored a commentary published in The New York Times yesterday (the 5th). They said that, based on a large number of documents recently uncovered by them at the Institute of National Security, GoLaxy was able to combine generative artificial intelligence with vast databases of personal information. The system could continuously mine data from social media platforms, build dynamic psychological profiles, and tailor content according to individuals’ values, beliefs, emotional tendencies and vulnerabilities.
生成式人工智能（Generative AI）的急速發展，使環球資訊戰進入一個全新時代。美國有專門研究國際安全事務的學者表示，中國科技公司「中科天璣」已升級為這方面的「領先者」，2020年在《國安法》實施後壓制香港反對聲音，2024年再出招干預台灣大選，目前更準備進軍美國，對全球民主制度構成前所未有的威脅。
高德斯坦（Brett J. Goldstein）與賓森（Brett V. Benson）是范登堡大學（Vanderbilt University）教授，兩人昨（5日）在《紐約時報》聯署發表評論文章指，根據他們在國家安全研究所（Institute of National Security）最近發現的大量文件，「中科天璣」（GoLaxy）能夠將生成式人工智能，與龐大的個人數據庫結合，系統能不斷從社交平台挖掘數據，建立動態的心理輪廓，並根據個人的價值觀、信念、情緒傾向與弱點，度身打造專屬內容。
30 Mar 2023明報#6237
警倡52億換通訊系統 前線可傳影畫 採用5G 另建影像平台AI辨疑人 議員促訂使用守則
Police have proposed spending about 5.2 billion to build next-generation communications system infrastructure using a 5G mobile broadband network, enabling frontline officers to transmit voice, video and images in real time while on duty. The system would replace the current terrestrial trunked radio technology, known as TETRA, which can only transmit voice; its cost includes the purchase of about 33,000 radio handsets. Police have also proposed spending about 390 million to build a digital image-processing platform to replace the use of DVD discs to back up video evidence. The platform would include artificial intelligence (AI) image-analysis tools to identify suspicious people and vehicles, enhancing police investigative and intelligence-analysis capabilities in various fraud cases and cases involving national security and public safety.
警方建議斥資約52億元，建立新一代通訊系統基礎設施，以流動寬頻5G網絡，讓前線警員執勤時可實時傳遞聲、影、畫等信息，取代目前只能用作語音傳輸的「地面中繼式無線電」（即TETRA）技術；通訊系統的費用包括添置約3.3萬個無線電對講機。警方另建議斥資約3.9億元建構數碼影像處理平台，取代以DVD光碟備份影片證據；並配置人工智能（AI）影像分析工具，辨識可疑人物及車等，提升警方在各類騙案、涉及國家安全及公眾安全案件的偵查及情報分析能力。
Translations are supplied by the source workbook. Consult the original reporting for context.
