The proposal, outlined in a draft circular, comes as digital transactions surge, particularly during the Lunar New Year period.
Under the draft, banks and e wallet providers must inform customers before deploying AI powered tools such as virtual assistants, automated hotlines and chatbots.
They are also required to disclose when AI is used for emotion recognition or biometric classification, and to clearly label any AI generated content, including images, audio and video.
The draft prohibits the use of AI to exploit customer vulnerabilities related to age, disability or financial hardship in marketing financial products.
Customers will have the right to lodge complaints over AI based decisions, with banks obliged to assign human staff to review such cases.
The regulations are expected to take effect from March, with existing AI systems given until September 2027 to comply.
The central bank said the framework aims to manage risks from the rapid adoption of AI, ensure fair treatment of customers and protect vulnerable groups.
Vietnamese banks have increasingly applied AI in areas such as eKYC, credit scoring, loan appraisal and fraud detection, while the rise of generative AI in customer service has raised concerns over data security, accuracy and privacy.