新的人工智慧安全威脅出現：惡意軟體蠕蟲和 Deepfake 聲音帶來挑戰

研究人員已經發現了 ChatGPT 等人工智慧工具帶來的潛在安全威脅。 Morris II 等惡意軟體蠕蟲可以利用人工智慧架構，在使用者不知情的情況下複製惡意提示來危害系統。為了防止這些威脅，用戶應謹慎對待未知的電子郵件和鏈接，投資防毒軟體，並使用強密碼、系統更新和有限的文件共享。此外，OpenAI 的新語音引擎可重現人類聲音，引發了人們對惡意行為者潛在利用的擔憂。

Novel Security Threats Emerge with the Rise of Generative Artificial Intelligence

隨著生成人工智慧的興起，出現了新的安全威脅

Recent advancements in artificial intelligence (AI) have propelled Generative AI tools, such as OpenAI's ChatGPT and Google's Gemini, to the forefront of technological innovation. While these tools hold immense promise for revolutionizing various industries, researchers caution that inherent security threats loom large, potentially jeopardizing their safe and widespread adoption.

人工智慧 (AI) 的最新進展將 OpenAI 的 ChatGPT 和 Google 的 Gemini 等生成式 AI 工具推向了技術創新的前沿。儘管這些工具為各行業帶來革命性的巨大希望，但研究人員警告說，固有的安全威脅迫在眉睫，可能會危及它們的安全和廣泛採用。

Malware Worms: A Looming Threat to Generative AI

惡意軟體蠕蟲：對產生人工智慧的迫在眉睫的威脅

In a groundbreaking study, researchers have uncovered a critical vulnerability in the architecture of Generative AI systems. This vulnerability, exploited by a sophisticated malware worm aptly named Morris II, could spell disaster for unsuspecting users.

在一項開創性的研究中，研究人員發現了生成式人工智慧系統架構中的關鍵漏洞。這個漏洞被一種名為 Morris II 的複雜惡意軟體蠕蟲所利用，可能會給毫無戒心的用戶帶來災難。

Similar to the infamous Morris worm of 1988, which crippled a staggering 10% of internet-connected computers, Morris II possesses the ability to self-replicate and spread relentlessly throughout Generative AI systems. Its insidious nature lies in its ability to bypass traditional security measures by compromising prompts, the textual instructions that guide GenAI's operations.

與 1988 年臭名昭著的莫里斯蠕蟲病毒（導致 10% 的聯網電腦癱瘓）相似，莫里斯 II 擁有自我複製的能力，並在整個生成人工智慧系統中無情地傳播。它的陰險本質在於它能夠透過損害提示（指導 GenAI 操作的文字指令）來繞過傳統安全措施。

Morris II manipulates prompts, transforming them into malicious directives that entice Generative AI platforms to perform destructive actions unbeknownst to the user or the system itself. This stealthy approach allows the worm to wreak havoc without triggering any red flags.

Morris II 操縱提示，將其轉換為惡意指令，誘使生成式 AI 平台在使用者或系統本身不知情的情況下執行破壞性操作。這種隱密的方法使蠕蟲能夠造成嚴重破壞，而不會觸發任何危險信號。

Safeguarding Against Malware Threats

防範惡意軟體威脅

In light of these alarming findings, experts urge Generative AI users to exercise heightened vigilance against suspicious emails and links originating from unknown or untrustworthy sources. Additionally, investing in robust antivirus software capable of detecting and eliminating malware, including these elusive computer worms, is strongly recommended.

鑑於這些令人震驚的發現，專家敦促生成式人工智慧用戶對來自未知或不可信來源的可疑電子郵件和連結保持高度警惕。此外，強烈建議投資購買能夠偵測並消除惡意軟體（包括這些難以捉摸的電腦蠕蟲）的強大防毒軟體。

Implementing strong password protection, regularly updating systems, and limiting file-sharing further bolster defenses against malware attacks. By adhering to these precautionary measures, users can significantly reduce the risk of malware infiltration.

實施強大的密碼保護、定期更新系統和限制檔案共享可以進一步增強對惡意軟體攻擊的防禦。透過遵守這些預防措施，使用者可以顯著降低惡意軟體滲透的風險。

Voice Engine: A New Frontier with Potential Security Implications

語音引擎：具有潛在安全影響的新領域

Amidst the concerns surrounding malware threats, OpenAI has unveiled a groundbreaking new tool called Voice Engine. This innovative technology leverages text input and a mere 15-second voice sample to recreate an individual's voice with remarkable accuracy.

鑑於對惡意軟體威脅的擔憂，OpenAI 推出了一款名為「語音引擎」的突破性新工具。這項創新技術利用文字輸入和僅 15 秒的語音樣本，以極高的準確性重新創建個人的語音。

While Voice Engine holds immense potential for diverse applications, researchers caution that its very nature could inadvertently become a tool for nefarious actors. The GenAI model underpinning Voice Engine could potentially be exploited to impersonate voices, leading to malicious activities.

雖然語音引擎在各種應用中具有巨大的潛力，但研究人員警告說，其本質可能會無意中成為不法分子的工具。支援語音引擎的 GenAI 模型可能會被用來模仿語音，導致惡意活動。

As Voice Engine transitions from its current testing phase to widespread availability, it is imperative that developers prioritize security measures to mitigate its potential misuse.

隨著語音引擎從目前的測試階段過渡到廣泛的可用性，開發人員必須優先考慮安全措施，以減少其潛在的誤用。

Regulatory Oversight and the Future of Generative AI

監管監督和生成人工智慧的未來

The emerging security threats associated with Generative AI have sparked concerns among regulatory bodies worldwide. Recognizing the potential for misuse, they are actively exploring frameworks to ensure the safe and ethical development and deployment of these technologies.

與生成人工智慧相關的新興安全威脅引發了全球監管機構的擔憂。認識到濫用的可能性，他們正在積極探索框架，以確保這些技術的安全和道德的開發和部署。

Striking a balance between innovation and security is paramount, ensuring that the transformative power of Generative AI can be harnessed without compromising user safety. As the field continues to evolve at a rapid pace, ongoing research and collaboration are essential to address the evolving security landscape and pave the way for a responsible and secure future for Generative AI.

在創新和安全之間取得平衡至關重要，確保在不損害用戶安全的情況下利用生成式人工智慧的變革力量。隨著該領域繼續快速發展，持續的研究和協作對於解決不斷變化的安全形勢至關重要，並為產生人工智慧的負責任和安全的未來鋪平道路。