周鴻祎：龍蝦很快就能“一鍵安裝”

近日，能替人干活的。全國兩會期間，多位代表、委員發表看法。

An AI agent known as "OpenClaw", nicknamed "lobster" by users for its ability to perform tasks autonomously, has gone viral online and drawn attention during the ongoing two sessions.

全國政協委員、360集團創始人周鴻祎表示，“龍蝦”是一個非常好的概念，把原來看不見摸不著的云上軟件，變成了每個人在電腦里養的專屬助手，而且操縱起來非常方便。

他同時指出，“龍蝦”現在還在早期，配置“龍蝦”對普通人來說是一件“非常難”的事?！拔覀兒芸鞎l一個一鍵安裝的版本，讓每個人養‘龍蝦’都很方便?！?/p>

Zhou Hongyi, a member of the 14th National Committee of the Chinese People's Political Consultative Conference and founder of 360 Security Group, said the technology turns cloud-based software into a personal assistant users can "raise" on their computers.

He added that it remains difficult for ordinary users to configure, but a one-click installation version is expected to be released soon.

全國政協委員、中國工程院院士王堅8日在接受采訪時表示，OpenClaw會很快便宜下來并普及，任何行業內的人都不會沒有看到它的存在。

Wang Jian, a member of the CPPCC National Committee and an academician of the Chinese Academy of Engineering, said such tools are likely to become cheaper and more widely adopted.

“‘養龍蝦’的走紅并非孤立現象，智能工具正快速進入大眾視野?！比珖f委員、中國科學院計算技術研究所研究員張云泉說，“這將對未來的工作方式乃至邏輯產生深遠影響”。張云泉說，過去需要人工完成的簡單重復任務，將逐步被智能體替代，人得以把更多精力投入創意設計、復雜決策等高價值工作。這也對從業者能力提出了更高要求。

張云泉表示，當前不少工業企業并非直接訓練大型模型，而是開發各類智能體，逐步替代生產流程中可智能化的環節，推動流程再造。在個人應用領域，數字員工乃至“一人公司”（OPC）等新模式正迅猛發展。

Zhang Yunquan, also a member of the CPPCC National Committee and a researcher at the Institute of Computing Technology of the Chinese Academy of Sciences, said the trend shows intelligent tools are rapidly entering public life.

He said intelligent agents may increasingly handle simple tasks, while companies are using them to automate production and explore models such as digital employees and "one-person companies".

安全風險仍是業內最關心的問題

近期，工業和信息化部網絡安全威脅和漏洞信息共享平臺監測發現OpenClaw開源AI智能體部分實例在默認或不當配置情況下存在較高安全風險，極易引發網絡攻擊、信息泄露等安全問題。

China's Ministry of Industry and Information Technology has warned that some instances of the open-source AI agent may pose security risks under default or improper configurations.

據《每日經濟新聞》報道，澳大利亞網絡安全公司Dvuln證明了這種風險，該公司發現OpenClaw存在漏洞，攻擊者可借此獲取用戶數月內的私人消息、賬戶憑證、API密鑰等敏感信息，一旦被黑客入侵，“一秒就可以搬空”。

OpenClaw存在的這種風險已引發“防御性關?！?。目前，韓國數家科技巨頭已正式下達禁令，限制員工在辦公設備上使用OpenClaw。一位行業內部人士透露，這并非不信任AI，而是為了“徹底封鎖內部機密被用于訓練外部模型的可能性”。

Australian cybersecurity firm Dvuln found vulnerabilities that could allow attackers to access users' private messages, account credentials and API keys. The risks have prompted several South Korean tech companies to ban the tool on office devices to prevent potential data leaks.

來源：中國青年報 中新經緯 央視新聞 保密觀

跟著China Daily

精讀英語新聞

“無痛”學英語，每天20分鐘就夠！