TAIPEI (TVBS News) — China's low-cost artificial intelligence (AI) model DeepSeek has captured global attention since its January release. However, the surge in AI has sparked cybersecurity concerns, leading Taiwan's Executive Yuan to ban DeepSeek's services in public sectors to safeguard national information security.
Ku Chen-hao (顧振豪), deputy director of the Institute for Information Industry's (資策會) Information Security Center (資安所), has raised alarms about the risks of data inaccuracies and personal information leaks in AI processes. Nations such as the United States, Japan, and India have implemented protective measures, underscoring the global emphasis on balancing innovation with security in AI technology.
In comparison to OpenAI, DeepSeek has impressed industry experts with its reduced training time and costs, instilling confidence in the future of AI applications. Within a month of its release, DeepSeek unveiled two models, V3 and R1, claiming training costs of less than US$6 million each—approximately one-tenth of the expenses incurred by other tech giants.
DeepSeek's pricing strategy further distinguishes it from its competitors. The R1 service charges US$0.55 per million input tokens and US$2.19 per million output tokens, whereas OpenAI's latest o1 model charges US$15 and US$60, respectively.
