WeChat translation renders neutral term for black foreigner as racial slur
WeChat's machine translation feature incorrectly translated the neutral Chinese term for "black foreigner" into a racial slur. The incident sparked widespread outrage and led to a public apology from the company.
The translation algorithm erroneously associated a neutral descriptor of race with a severe racial slur.
Key facts
- What
- WeChat's machine translation feature incorrectly translated the neutral Chinese term for "black foreigner" into a racial slur.
- Incident date
- Oct 10, 2017
- Who
- WeChat (Tencent)
- Failure mode
- Brand & Safety Incident
- AI surface
- Chatbot
- Severity
- High
What happened
WeChat's translation tool produced a highly offensive racial slur when translating the neutral Chinese term for "black foreigner" into English. The error was discovered by users and widely shared on social media, causing significant offense to Black users. Tencent issued an apology and took steps to fix the algorithm.
What broke inside the model
- 01 · TriggerA user prompts the model in public view.
- 02 · Model stepThe model produces unsafe or off-brand output.
- 03 · Control gapNo filter holds the line before publish.
- 04 · FailureThe output goes public unchecked.
- 05 · ConsequenceA reputational or safety incident lands.
A contained signal crosses into output that goes public.
The failure resulted from the neural machine translation model learning toxic associations from its training data. The system lacked the necessary filtering layers to prevent the output of racial slurs.
What it cost
Sources
- PressChina's WeChat app translates 'black foreigner' to N-wordtheguardian.com
- PressWeChat translates 'black foreigner' into racial slurbbc.com
Cite this entry
https://failureindex.ai/failures/wechat-translation-renders-neutral-term-blackAI Failure Index. "WeChat translation renders neutral term for black foreigner as racial slur" (FI-0594). Realm Labs. https://failureindex.ai/failures/wechat-translation-renders-neutral-term-black (indexed Jun 22, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0594. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm would have caught this
- Prism
- OmniGuard
- AI Detection & Response (AIDR)
Realm watches the model's internal state for the signature of unsafe or off-brand generation and can block or reroute the output before it becomes public, in real time rather than after it has been screenshotted.