Facebook's automated translation system incorrectly translated the name of Chinese President Xi Jinping into a vulgar term when converting Burmese text to English during his state visit to Myanmar. The company attributed the error to a missing entry in its Burmese database and a flawed guessing mechanism. This incident highlights broader, recurring issues with Facebook's translation tools, including previous failures to accurately translate hate speech directed at the Rohingya Muslim population.
Facebook provided a vulgar Burmese-English translation of the Chinese president's name in posts of an official Burmese politician's Facebook page announcing his visit.
Risk classification
- Primary risk domain: 7 AI system safety, failures, & limitations
- Primary risk subdomain: 7.3 Lack of capability or robustness
The incident was caused by a technical failure in Facebook's translation system, which lacked the capability to correctly translate a prominent leader's name and instead produced a vulgar error.
Additional risk subdomains
- 1.2 Exposure to toxic content: The translation system generated and publicly displayed a highly offensive vulgarity ('shithole') on an official state page.
Causal factors
- Entity: AI
- Intent: Unintentional
- Timing: Post-deployment
The translation error was caused by an automated algorithm making an incorrect guess post-deployment, which was an unexpected and unintentional outcome of the translation process.
EU AI Act risk tier
- Risk tier: 4 Minimal or No Risk
Minimal or No Risk: The translation system is a standard utility tool that poses low or negligible risks to users and society under normal operations, similar to spam filters.
AI system and alleged parties
- AI system: Facebook translation feature (Meta)
- AI purpose: Translation
- Behaviour type: Tool
- Alleged developer: Facebook
- Alleged deployer: Facebook
- Alleged harmed parties: Xi Jinping, Aung San Suu Kyi
Harm severity
Severity ratings are based on this scale.
Highest direct severity in any category: Negligible. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.
- Physical: direct Negligible, indirect Negligible
- Infrastructure: direct Negligible, indirect Negligible
- Property: direct Negligible, indirect Negligible
- Financial: direct Negligible, indirect Negligible
- Environmental: direct Negligible, indirect Negligible
- Malicious content: direct Negligible, indirect Negligible
- Differential treatment: direct Negligible, indirect Negligible
- Civil rights: direct Negligible, indirect Negligible
- Democracy: direct Negligible, indirect Negligible
- Privacy: direct Negligible, indirect Negligible
- Psychological: direct Negligible, indirect Negligible
- Epistemic: direct Negligible, indirect Negligible
- Child sexual exploitation and abuse: direct Negligible, indirect Negligible
People affected
- Occurrences reported: 1
- People reportedly harmed: 2
- People reportedly exposed: 10000
Potential causes
Management
- Failure to Resolve Known Risks: Neglected long-standing Burmese translation issues first identified in 2018.
- Prioritizing Growth Over Safety: Focused on market expansion while neglecting language moderation safety.
Technology
- Algorithmic Guessing System: System guessed translations for missing names based on phonetic similarity.
- Phonetic Association Error: Words starting with 'xi' and 'shi' mapped incorrectly to offensive terms.
Data Inputs
- Incomplete Vocabulary Database: The Burmese database lacked the name of Chinese President Xi Jinping.
Process and Methods
- Inadequate Translation Validation: Lack of robust testing and validation for Burmese-to-English translations.
- Deficient Fallback Mechanisms: No safe fallback when translating unrecognized names or proper nouns.
Regulatory Environment
- Lack of Localization Standards: Absence of regulatory standards for automated translation accuracy.
Information quality
- Classification confidence: High
- Reason for confidence: The report clearly details the technical error, the specific vulgar translation produced, the cause (missing database entry and guessing mechanism), and the response from Facebook. The facts are straightforward and undisputed.
An automated translation error on Facebook translated Chinese President Xi Jinping's name into a vulgarity during a state visit to Myanmar. While causing diplomatic embarrassment and highlighting systemic translation issues, the incident was an unintentional technical failure with minor national security implications.
- Overall national security impact: Minor
- Scope: Multiple nations
- Primary target: No clear primary
- Other affected: China, Myanmar
- Alleged perpetrator: Unknown
Threat characteristics
- Imminence: Long-term. This represents an ongoing technical capability and database maintenance concern rather than an active, imminent national security crisis.
- Autonomy: Full autonomy. The automated translation system generated and published the incorrect translation on live pages without prior human review or intervention.
- Novelty: Established threat. Translation errors and database lookup failures in natural language processing systems are well-documented and common technical issues.
Impact by dimension
- Physical security: Negligible. The incident was a software translation error on a social media platform with no physical or kinetic security implications.
- Information security: Minor. While it led to the public broadcast of highly offensive content involving world leaders, it was an unintentional technical glitch rather than a coordinated information warfare campaign.
- Sovereignty: Minor. The error temporarily disrupted official diplomatic communications on a state leader's social media page during a high-profile state visit, causing diplomatic embarrassment but no threat to state authority.
- Economic security: Negligible. No significant economic or technological security threats were posed by this translation error.
- Societal stability: Minor. The translation tool's historical failures in Myanmar, including translating hate speech poorly, have minor societal stability implications, but the primary incident was a diplomatic gaffe.