2024年5月8日，新加坡內政部長尚穆根先生在國會書面答覆蘇慧敏議員政府如何打擊「深偽」圖像或視頻實施犯罪的問題。
以下內容為新加坡眼根據國會英文資料翻譯整理：
針對利用「深偽」圖像或視頻實施犯罪的罪犯的措施
21 蘇慧敏女士詢問內政部長是否有足夠的保護措施來防止犯罪者似乎越來越多地利用「深偽」圖像或視頻實施各種犯罪（包括威脅和勒索）的犯罪者。
尚穆根先生：內政部一直在與新加坡通訊和信息部、資訊通信媒體發展局（IMDA）和網絡安全局合作，共同應對涉及深偽或以其他方式由人工智慧協助的犯罪。
《網絡犯罪危害法令》（OCHA）允許政府向網絡服務提供商發出指示，以防止網絡犯罪活動與新加坡用戶互動或接觸。其中包括用於實施威脅和勒索等犯罪行為的深偽圖像或視頻。
《網絡犯罪危害法令》還允許我們要求指定網絡服務提供商制定措施，以阻止詐騙和惡意網絡活動，包括由"深偽"提供便利的活動。這些措施將在今年推出的《網際網路行為準則》中加以規定。
此外，《防範網絡假信息和網絡操縱法》（POFMA）使政府能夠在符合公共利益的情況下發布更正，並標記對公眾具有欺騙性或誤導性的「深偽」產品。該法令還允許政府發布禁用指令，限制新加坡最終用戶訪問此類「深偽」產品。
政府還將繼續加強我們的技術能力以應對這些威脅。例如，新加坡警察部隊正在與內政部科技局合作，開發檢測人工智慧生成的音頻和視頻的技術。由通訊及新聞部發起，並由科技研究局主辦的網絡安全先進技術中心也在開發工具，以檢測有害網絡內容（包括「深偽」）。儘管如此，這是一個嚴峻的挑戰，技術會不斷進步，很難解決所有問題。
以下是英文質詢內容：
MEASURES AGAINST PERPETRATORS WHO ADOPT DEEPFAKE IMAGES OR VIDEOS TO COMMIT OFFENCES
21Ms Hany Soh asked the Minister for Home Affairs whether there are sufficient protective measures against perpetrators who appear to be increasingly adopting deepfake images or videos to commit various offences, including making of threats and extortion.
Mr K Shanmugam: The Ministry of Home Affairs has been working with the Ministry of Communications and Information, the Infocomm Media Development Authority and the Cyber Security Agency of Singapore to address crimes which involve deepfakes, or which are otherwise facilitated by Artificial Intelligence.
The Online Criminal Harms Act (OCHA), allows the Government to issue directions to online service providers to prevent online criminal activities from interacting with or reaching Singapore users. These include deepfake images or videos used to commit offences, such as the making of threats and extortion.
OCHA also allows us to require providers of designated online services to put in place measures to disrupt scams and malicious cyber activities, including those facilitated by deepfakes. These measures will be stipulated in the Codes of Practice which will be introduced this year.
In addition, the Protection from Online Falsehoods and Manipulation Act (POFMA), enables the Government to issue corrections and label deepfakes that are deceptive or misleading to the public, where there is public interest to do so. POFMA also allows the Government to issue disabling directions to restrict access by Singapore end-users to such deepfakes.
The Government will also continue to strengthen our technical capabilities to deal with these threats. For example, the Singapore Police Force is working with the Home Team Science and Technology Agency to develop technologies to detect AI-generated audios and videos. The Centre for Advanced Technologies in Online Safety, an initiative by the Ministry of Communications and Information and hosted by the Agency for Science, Technology and Research, is also building tools to detect harmful online content, including deepfakes. That said, this is a serious challenge and technologies will continuously evolve. It is difficult to pick up all cases.
HQ丨編輯
HQ丨編審
新加坡國會丨來源
