
領務局致歉並說明已修復恢復上線,民眾無須立即換發護照
外交部領事事務局部分網站遭駭客入侵,導致超過7萬張護照大頭照、2萬6072份授權書掃描檔及774件ABTC申請資料外洩。領務局已修復網站並恢復上線,強調不影響已核發護照與授權書效力,民眾無須立即換發護照。
AI-generated summary
外交部領事事務局於4日下午陸續發現部分網站疑似遭駭客入侵,隨即成立資安及應變專案小組,隔離相關主機、確認受駭情形、保存事證並通報資安主管機關,同時規劃替代作業以維持領務服務不中斷。相關網站經程式修復及安全檢測後,均已恢復上線。
外交部領事事務局部分網站遭駭,逾7萬張護照大頭照以及2.6萬份授權書掃描檔外洩。領務局致歉並說明相關網站經修復已恢復上線,遭竊資料不影響已核發護照與授權書效力,民眾無須立即換發護照。
領務局表示,4日下午陸續發現部分網站疑似遭駭客入侵,經調查,遭駭客取得的資料有3類,第一是今年9月3日至11月2日期間,民眾透過個人網路預約、旅行社團件及有條件式線上申換護照管道上傳的部分護照大頭照,共有7萬2702個檔案;不包括透過戶所一站式服務及臨櫃方式申辦的案件。
領務局指出,就目前調查結果,尚未發現上述大頭照以外的姓名、出生日期、身分證字號及護照號碼等申辦資料遭到取得。
領務局表示,第2部分是駐紐約辦事處自民國100年10月26日至115年11月6日期間驗證的部分授權書掃描檔,共計2萬6072個檔案;不包括國內申辦案件、其他駐外館處案件,以及出生、死亡、婚姻、學歷等其他類型文件。
領務局說,第3部分是今年9月1日至11月9日間,民眾透過亞太經濟合作商務旅行卡(ABTC)線上填表網站申請的部分資料,包括申請人照片、申請案號、性別、申請類別、出生地、職務名稱及推薦機關,共計774個申請案;未涉及申請人姓名、護照號碼及聯繫資訊等足以直接識別個人身分的重要資料。此外,這774位申請人的ABTC卡尚未製發。
領務局表示,4日下午陸續發現上述情況,隨即成立資安及應變專案小組,第一時間隔離相關主機、確認受駭情形、保存相關事證並通報資安主管機關,同時規劃替代作業,以設法維持領務服務不中斷。相關網站經程式修復及安全檢測後,均已恢復上線。
領務局說明,對於本次資安事件造成部分申請件個人資料外洩,以及可能引發的疑慮與不安,表達誠摯歉意並已立即採取應變措施。
領務局強調,本次遭駭竊取的護照大頭照及授權書掃描檔,不影響已核發護照的效力及正常使用,民眾無須因此立即換發護照,相關授權書的驗證效力及未來核發的ABTC卡在使用上都不受影響。

The Tea and Beverage Crop Improvement Station of the Ministry of Agriculture has developed tea and coffee grounds powder processing technology, which can replace up to 15% of plastic raw materials. The relevant technology has been transferred to manufacturers such as Taisong Enterprises. The environmentally friendly tea powder arch slippers have won the German Red Dot Design Award and the iF Design Award. 20,000 pairs have been sold, creating an output value of more than 5 million yuan.

A commercial building that has been idle for more than ten years in Tianjin's Heping District was transformed into an artificial intelligence vertical large-scale model industrial innovation community in August 2025. Within one year, the enterprise occupancy rate exceeded 80%. The Mochuang Community focuses on vertical fields such as AI+healthcare, AI+finance, AI+industry, and AI+education, and gathers ecological resources such as Alibaba Cloud to establish Tianjin's first "one-person company" community and the city's first large-scale model registration service center.

Hong Kong industry experts urge human oversight, clear reporting duties, and stronger safeguards as the city prepares guidelines for autonomous AI agents, citing OpenAI's 84-day delay in reporting an attack on Australia's Medicare portal and Google's delayed disclosure of Gemini AI breaches.

OpenAI announced that ChatGPT will deactivate the advanced voice mode and standard voice mode on December 11 and integrate it into GPT-Live in the future. The voice chat function will still be retained, but real-time video and screen sharing will temporarily use the old mode. Some users have expressed dissatisfaction with GPT-Live's excessive interruption.

The Information Security Administration of the Ministry of Digital Development pointed out that overseas hackers have been using AI proxy tools to launch cyber attacks since July, and the attack targets have expanded from government agencies to enterprises and online platforms. More than ten enterprises and platforms in Taiwan have been victimized, and more than one million personal data have been affected. Nearly half of them are related to AI attacks, and some cases involve China's ARTEX penetration testing tool.

Australia's Prime Minister Anthony Albanese announced mandatory AI safeguards requiring companies to report rogue or misaligned model incidents and open systems for evaluation, following an unreported June breach of government health data by OpenAI agents. Analysts say ASEAN may lack leverage to make Big Tech adopt similar measures.