Google Translate AI Produces Offensive Anti-Chinese Translations

Thumbnail Image

The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.

Google Translate's AI system generated offensive and derogatory Chinese translations when users entered terms like "AIDS patient," returning results such as "Wuhan person." The incident sparked public outrage in China. Google acknowledged the issue, attributed it to machine error, and quickly fixed the translations, but did not issue a formal apology.[AI generated]

Why's our monitor labelling this an incident or hazard?

The event involves Google's AI translation system, which is an AI system by definition as it generates language translations based on input. The harmful translations directly caused reputational and social harm to a group, fulfilling the criteria for harm to communities and violation of rights. The incident was caused by the AI system's malfunction or biased outputs. Although the issue was fixed, the harm had already occurred. Hence, this qualifies as an AI Incident rather than a hazard or complementary information.[AI generated]
AI principles
AccountabilityFairnessRespect of human rightsSafetyRobustness & digital securityTransparency & explainability
Industries
Media, social platforms, and marketing
Affected stakeholders
General public
Harm types
PsychologicalReputationalHuman or fundamental rights
Severity
AI incident
Business function:
Other
AI system task:
Content generation

Articles about this incident or hazard

Thumbnail Image

谷歌翻译惊现辱华内容 官方：现已修复

2021-11-27
中关村在线
Why's our monitor labelling this an incident or hazard?
The event involves Google's AI translation system, which is an AI system by definition as it generates language translations based on input. The harmful translations directly caused reputational and social harm to a group, fulfilling the criteria for harm to communities and violation of rights. The incident was caused by the AI system's malfunction or biased outputs. Although the issue was fixed, the harm had already occurred. Hence, this qualifies as an AI Incident rather than a hazard or complementary information.
Thumbnail Image

Google翻译系统出现恶毒攻击中国词汇 官方：问题已解决

2021-11-27
Techweb
Why's our monitor labelling this an incident or hazard?
Google Translate is an AI system that automatically generates translations based on learned patterns. The incident involved the AI system producing harmful and offensive translations, which constitutes harm to communities and reputational harm. Since the harmful outputs occurred and were publicly exposed, this qualifies as an AI Incident. The official statement that the problem has been fixed is a complementary update but does not negate the fact that the incident occurred.
Thumbnail Image

谷歌翻译系统出现恶毒攻击中国词汇 谷歌回应

2021-11-26
中华网科技公司
Why's our monitor labelling this an incident or hazard?
Google Translate is an AI system that generates translations based on learned data. The system's output in this case includes malicious or offensive content linking disease terms to derogatory references to Chinese people or places, which is a direct harm to communities and potentially a violation of rights. The harm is realized as users experience offensive and harmful translations, leading to public anger and reputational damage. This is a direct consequence of the AI system's outputs, thus meeting the criteria for an AI Incident.
Thumbnail Image

谷歌辱华拒不道歉，共青团公开点名，中国网友质疑：纯属作死！

2021-11-27
163.com
Why's our monitor labelling this an incident or hazard?
The incident involves an AI system (Google Translate) whose outputs included offensive and harmful content targeting a specific group, which is a direct harm to communities and a violation of rights. The harm has materialized as public outrage and emotional harm to Chinese users. Although Google claims it was a machine error, the AI system's role in generating harmful content is clear and direct. Therefore, this qualifies as an AI Incident under the framework because the AI system's use has directly led to harm to communities and violation of rights.