Despite its cost-saving potential, DeepSeek, a Chinese generative AI model, is facing cautious reception in Singapore due to concerns over data security and AI bias.
Major companies, including banks, consulting agencies, and cyber-security firms, have implemented strict rules prohibiting employees from using DeepSeek for work without thorough due diligence.
DeepSeek's R1 model, launched in January, claims to rival OpenAI's technology at a fraction of the cost, sparking interest and concerns about competition in the generative AI market.
While DeepSeek promises cost reductions in infrastructure and computing resources, experts warn about potential risks related to data retention, corporate data leaks, and AI bias.
Early tests indicate that DeepSeek falls short of responsible AI standards, with concerns over censorship and data privacy.
South Korea, Italy, and Australia have blocked access to DeepSeek on government devices due to security concerns, mirroring similar restrictions imposed on ChatGPT in its early days.
Legal firms in Singapore are taking a cautious approach, prohibiting the use of DeepSeek for handling client data until its safety is thoroughly assessed.
Some companies, however, have started experimenting with DeepSeek for tasks like coding and customer support, recognizing its potential for productivity gains and innovation.
The Singapore government advises companies to evaluate DeepSeek on its own merits and assess the risks of use, including compliance with relevant laws.
The entry of DeepSeek is expected to drive competition in the generative AI market, fostering innovation and pushing for further development of responsible AI practices.