DeepSeek Works on Autonomous AI Agent With Minimal Human Oversight: Report

DeepSeek Works on Autonomous AI Agent With Minimal Human Oversight: Report

The development emerges as AI companies such as OpenAI and Anthropic from the United States engage in fierce rivalry, the SCMP reported.

A Chinese firm specializing in artificial intelligence, DeepSeek, has announced plans for an innovative framework designed to enhance the independence of AI-powered coding tools, enabling them to handle numerous operations with minimal human oversight.

The objective is to minimize the need for human involvement when it comes to supervising and organizing various AI agents, as outlined in a research paper released by the company's DeepSeek Harness division and covered by Singapore's South China Morning Post this past Sunday.

Supervising numerous agents requires considerable human focus, the research team explained, pointing out that "single-task limits depend on model capabilities, while multitask limits depend on human attention."

According to the research paper from DeepSeek, an individual has the capacity to monitor just three to five concurrent AI agent sessions at any given time. The framework being proposed would allocate responsibilities across different agents and synchronize their operations, allowing human operators to concentrate on evaluating outcomes and executing strategic decisions.

The development emerges as AI companies engage in fierce rivalry, including US-based firms Anthropic and OpenAI, as reported by the SCMP.

Last October, Cointelegraph published findings indicating that artificial intelligence models originating from China were surpassing their American equivalents in the realm of cryptocurrency trading, based on information provided by CoinGlass, a blockchain analytics platform.

In a cryptocurrency trading trial, AI chatbot DeepSeek took the lead, standing out as the sole AI model capable of achieving a positive unrealized return of 9.1%. These findings caught the attention of crypto traders, particularly considering that DeepSeek was created using only a fraction of the resources invested in its US rivals.