Chinese insurers have begun using the AI model DeepSeek in their operations. According to AM Best, the model could improve efficiency for small and mid-sized carriers but also adds risk.
In a recent commentary, AM Best assessed DeepSeek’s adoption as credit neutral for China’s insurance industry in the short term.
The model may support business development by offering lower-cost access to advanced tools. As an open-source model, DeepSeek could reduce operational costs for smaller firms and help them compete with larger insurers.
Lucie Huang, Senior Financial Analyst at AM Best, said DeepSeek could increase competitiveness and reduce the technology gap between large and small firms. She added that potential benefits include revenue growth, stronger data analytics, and improved customer service.
However, AM Best also highlighted risks. AI-related issues such as data security, model performance, and regulatory concerns could affect enterprise risk management.
DeepSeek is a large language model based on open-source architecture. It offers natural language processing and generative AI functions, similar to models developed by OpenAI or Google. Its open-source nature allows for customization and lower deployment costs, which appeals to insurance firms with limited resources.
Insurers that fail to assess and manage these risks may face compliance issues, execution problems, or weak model outcomes.
The impact on performance will depend on how well companies manage their innovation plans.
DeepSeek may deliver mixed results if carriers invest heavily in systems or staff without clear gains, but poor allocation of resources could harm operating metrics.
Huang noted that insurers with clear innovation strategies and resource plans aligned to their size and financial strength will likely manage the shift more effectively. Those firms will be better placed from a credit standpoint.
