While some organizations, like OCBC Bank, are restricting the use of external chatbots, others are exploring DeepSeek for productivity gains, recognizing its potential for cost savings.

Despite its cost-saving potential, companies in Singapore are hesitant to adopt DeepSeek due to concerns about data security and potential AI biases.

The Singapore Ministry of Digital Development and Information has advised companies to independently evaluate the merits and risks of using products like DeepSeek.

Legal firms in Singapore are taking a conservative approach, prohibiting the use of generative AI tools for client data until they demonstrate safety and compliance with regulations.

Experts believe that DeepSeek's entry into the market may enhance competition in the generative AI sector, potentially fostering innovation despite the associated risks.

DeepSeek's R1 model, launched in January 2025, has gained attention for its cost-effectiveness, with training costs significantly lower than those of major competitors, at S$7.6 million.

Ai Palette estimates that DeepSeek could reduce infrastructure costs for companies by 40% to 60%, particularly in high-end computing needs.

Early evaluations indicate that DeepSeek may not meet certain responsible AI standards, including issues with answering sensitive questions and potential biases.

Major firms, including banks and consulting agencies, have implemented strict rules against using generative AI tools like DeepSeek without thorough assessment due to concerns regarding due diligence.

Boston Consulting Group has reported increased interest in DeepSeek, highlighting its open-source nature and cost-efficiency, while also advising caution regarding its deployment.

Concerns regarding generative AI tools include data retention, corporate data leaks, and compliance with responsible AI standards.

Governments in South Korea, Italy, and Australia have restricted access to DeepSeek on government devices due to security concerns, reminiscent of earlier restrictions on ChatGPT.