AI is already part of daily work in Taiwan's legislature, and staff and academics warn it could quietly shape political values and carry Chinese bias.

• Lawmakers' offices now use AI to draft speeches, build slides and graphics, and analyze laws.
• Staff say AI is useful for sorting information, but it cannot decide moral questions like euthanasia, the death penalty, or how to treat the poor.
• One office director warns that letting AI tailor a message to whatever the audience wants to hear is not political judgment, just empty wordsmithing.
• A legislative aide calls AI a "new god" people treat as an authority, making humanities training and critical thinking more valuable, not less.
• Security expert Ho Cheng-hui says the underlying model and training data shape the output, so users should tell AI to exclude Chinese sources.

Outlook: Expect pressure in Taiwan for training on how political staff use AI, and more attention to which models and data sources they rely on.