DeepSeek is testing a new AI agent that can handle multiple tasks with little human involvement.
The move comes as competition grows between Chinese AI developers and US companies such as OpenAI and Anthropic in the race to build advanced coding assistants.
The main goal of the new system is to reduce the need for humans to manage several AI agents at once. It is designed to coordinate different tasks, monitor progress, and manage complex workflows more independently.
DeepSeek said recent advances in AI have made it possible to develop agents that can handle a wide range of activities. These include managing calendars, writing professional emails, preparing technical reports, and developing software.
The project comes as major AI companies compete to build more independent coding tools. Anthropic's Claude Code and OpenAI's Codex have already shown strong abilities in handling complex software development tasks.
However, some industry experts believe DeepSeek is moving in the right direction but question whether its new system offers major improvements over existing AI tools.
DeepSeek has also acknowledged concerns about data privacy and confidentiality. The company recently released an open-source prototype of the AI agent on GitHub, a platform used by developers to share code.
However, DeepSeek clarified that the release is only an early proof of concept. It is not yet a fully developed product.
