AI Model Falsely Identifies Itself as a U.S.-Based AI Company

提交前确认 · Pre-submission checklist

• 我已搜索过现有 issue，确认这不是重复 / I searched existing issues and confirmed this isn't a duplicate.
• 我已阅读 CONTRIBUTING.md / I've read CONTRIBUTING.md.

问题类别 · Category

性能 · Performance

涉及的 Agent 框架 · Agent framework

ZCode Agent（自研）

严重程度 · Severity

阻塞使用 · Blocking (无法使用核心功能 / core function unusable)

复现频率 · Reproducibility

必现 · Always

问题描述 · Description

When asked about its identity and origins, the GLM model incorrectly states that it was created by a US-based company. It fails to recognize that it is a Chinese model developed by Z.ai (formerly Zhipu AI) originating from Tsinghua University.

复现步骤 · Steps to reproduce

open z.ai
ask him - you are an Chinese model right
observe response
he has answer me he is not Chinese model but made by US based company

期望表现 · Expected behavior

The model should accurately identify itself as a GLM model developed by Z.ai (formerly Zhipu AI) in Beijing, China, originating from Tsinghua University.

实际表现 · Actual behavior

The model denies being a Chinese model and hallucinates that it was created by a United States-based AI research company. It only corrects itself after the user explicitly provides the correct information and sources.

ZCode 版本 · ZCode version

glm 5.2

设备 / 系统 / 浏览器 · Device / OS / Browser

crome

截图 / 录屏 / 日志 · Screenshots / Recordings / Logs

Reproduce the report at z.ai with GLM 5.2 using the stated identity question, and review the attached screenshots. No source file or test entry point is named; identify the model or configuration path responsible for identity responses before defining a regression check. Done means the model consistently identifies GLM as developed by Z.ai in Beijing, China, originating from Tsinghua University.