Altman Builds Exclusive Platform and “AI Assistant to Be Fully Integrated into Work This Year”
Manhouse Emerges in China, Signaling Intense Competition… Security Concerns Rise,
Sam Altman, OpenAI CEO at X (formerly Twitter),
,
, “(Seoul=News1) Reporter Kim Min-seok = On the 11th (local time), OpenAI released the ‘Responses API,’ a software development platform for AI agents that go beyond the level of chatbots and perform tasks autonomously.”,
,
, ‘The Responses API supports developers and companies in using OpenAI models to create and establish AI agent software (Vertical AI) for performing specific tasks independently.’,
,
, “Responses API is expected to gradually replace the ‘Assistant API’ (supported until the first half of 2026). Although OpenAI introduced AI agents like ‘Operator’ in ChatGPT, it has only provided basic functions to paid subscribers.”,
,
, ‘Sam Altman, OpenAI CEO, announced the launch of Responses API on X (formerly Twitter) and commented, “Although there is bias against (AI agents), I believe that the model our team developed is among the best designed and most useful APIs.”‘,
,
, ‘Brad Lightcap, OpenAI COO, stated, “People can now interact with AI agents,” adding, “AI agents can refer to files, browse the web, and use computers independently.”‘,
,
,
OpenAI Web Browser AI Agent Operator,
,
, ‘Big tech companies view AI agent technology as moving beyond the experimental stage this year, becoming a key tool that enhances work efficiency and creates new value. There is an expectation that AI agents will become essential partners in daily life and work settings.’,
,
, ‘Altman also called this year “the year AI agents are fully integrated into work.”‘,
,
, “Indeed, in January this year, OpenAI introduced the ‘Operator,’ an AI agent based on a PC web browser. The Operator is capable of understanding and executing complex commands sequentially by interacting with what appears on the web browser screen.”,
,
, “Recently, it was reported that OpenAI plans to release a ‘PhD-level’ AI agent product with a monthly subscription fee of $20,000 (approximately 29 million KRW). The company is also considering launching a $2,000 (approx. 2.9 million KRW) monthly plan for high-income professionals and a $10,000 (approx. 14.42 million KRW) monthly plan for software development use.”,
,
,
Yichao Picci introducing AI Agent Manus AI in a video (X),
,
, “China’s AI startup Monica recently introduced ‘Manus AI,’ promoted as a general-purpose AI agent capable of autonomous judgment and action. However, from a performance perspective, its application is limited, and it frequently makes basic errors, highlighting the limitations of its initial version.”,
,
, ‘Currently, AI agents in the platform environment require access to a large amount of personal information and permissions to execute apps and complete tasks independently.’,
,
, ‘Security experts warn that granting AI agents broad control could invite critical threats such as hacker attacks and AI-specific errors or mistakes.’,
,
, ‘Altman has also expressed concern on his blog, stating, “Many of us fear losing control in an open-source manner and believe humans should retain more technical control,” and mentioned the potential use of AI by authoritarian governments to control citizens through mass surveillance and loss of autonomy.’