The Chinese artificial intelligence system DeepSeek is developing an advanced AI agent feature that can learn and improve over time, according to media reports.
DeepSeek’s upcoming AI agent features will be able to complete complex tasks for users with minimal input, Bloomberg reported Thursday. The system will also learn and grow as a user interacts with it, according to anonymous sources cited in reporting.
DeepSeek founder Liang Wenfeng is targeting the fourth quarter of this year for the planned release date for the AI agent features.
The launch follows a shift in the industry as AI companies look to create semi-autonomous agents that can perform intricate professional tasks without much supervision. In recent months, some of the biggest names in the industry — Anthropic, Microsoft and OpenAI — have launched versions of AI agents.
AI agents can do much more than carry out a search query, with most companies pitching the models as ways for businesses to better integrate AI into their everyday operations. AI agents can be used to automate complex workflows, analyze large amounts of data, debug code or provide custom experiences to customers.
The release of AI agent tools for DeepSeek could be the much-anticipated next step for the Chinese developer. While the release of DeepSeek’s R1 model in January shocked the world with its capabilities, the developer has released relatively few upgrades compared to its competitors.
