Bengali catches AI on the you it chooses: tui, tumi, and apni each carry a different closeness, and the model picks wrong. Its script handling can wander, and phrasing from Bangladesh and West Bengal is not interchangeable, though it often treats them as if they were.
Remote Language Data Jobs in North America 2026
539 remote jobs open to candidates in North America in Language Data. Updated daily.
Tagalog trips AI on the verb focus system and on the po and opo markers that carry respect. It handles the natural mix of Tagalog and English awkwardly and reaches for stiff, formal phrasing where real conversation code-switches freely, which a native speaker catches at once.
Russian trips AI on structure it cannot fake: six cases, verb aspect, and agreement that has to hold across a whole sentence.
AI produces textbook Italian that still feels off: it forces gender and number agreement mechanically, misses the formal Lei, and leans on wording that sounds translated rather than said.
German gives AI plenty to get wrong: it botches case endings, builds compound nouns no one would actually say, and misses when du should be Sie. Fluent-looking output can still feel mechanical, and only a native ear catches when Austrian or Swiss usage would read differently.
AI writes fluent-looking French that still slips on what natives feel automatically: it mishandles tu and vous, forces gender agreement in odd places, and reaches for wording that sounds translated rather than spoken.
Automated metrics can't judge whether Turkish text actually sounds right: tone, idiom, cultural nuance, and the thousand small choices a native speaker makes without thinking.
Automated metrics can't judge whether Indonesian text actually sounds right: tone, idiom, cultural nuance, and the thousand small choices a native speaker makes without thinking.
Automated metrics can't judge whether Arabic text actually sounds right: tone, idiom, cultural nuance, and the thousand small choices a native speaker makes without thinking.
Automated metrics can't judge whether Portuguese text actually sounds right: tone, idiom, cultural nuance, and the thousand small choices a native speaker makes without thinking.
Automated metrics can't judge whether Hindi text actually sounds right: tone, idiom, cultural nuance, and the thousand small choices a native speaker makes without thinking.
Automated metrics can't judge whether Spanish text actually sounds right: tone, idiom, cultural nuance, and the thousand small choices a native speaker makes without thinking.
Korean punishes the mistakes AI loves to make: it mixes speech levels, sliding between formal and casual within a single exchange, and drops or misplaces the particles that carry meaning. The honorifics a native speaker chooses without thinking are exactly where the model slips.
Vietnamese is unforgiving of the small errors AI makes constantly: it drops or misplaces the tone diacritics that change a word entirely, and it stumbles over the pronoun system, where the right term depends on age and relationship.
Thai gives AI little room: words run together without spaces, tone carries meaning, and the politeness particles shift with who is speaking.
AI can string together grammatical Mandarin and still read as foreign: it fumbles measure words, reaches for stiff textbook phrasing, and drops idioms like chengyu at the wrong pitch.
Japanese exposes AI quickly: it mishandles keigo, mixing polite, humble, and honorific speech in ways that would embarrass a real speaker, and it fumbles counters and particles.
1. Project type Code-switching recording is the practice of alternating between two languages within a single conversation.
We’re hiring remote freelancers for a Code-Switching Recording Project at AIxBlock. What you’ll do Record natural 2-person conversations that switch between two languages in the same dialogue.
(Native Japanese Writing Experts Only) We are seeking for native Japanese writing experts to contribute to the training of AI models aimed at enhancing writing capabilities.
(Native Korean Writing Experts Only) We are seeking for native Korean writing experts to contribute to the training of AI models aimed at enhancing writing capabilities.
(Native Tagalog Writing Experts Only) We are seeking for native Tagalog writing experts to contribute to the training of AI models aimed at enhancing writing capabilities.
Remote • English Voice Recording • Under 60 Minutes • Earn $15–$27+ Per Completed Task Your voice can help us better understand how people from different language backgrounds naturally speak and interact with voice technology.
What if your ear for detail and obsession with accuracy could directly shape how AI understands human speech? We're looking for Audio Transcription & Alignment Specialists in the Sydney area to clean up word-level transcripts inside a proprietary editor — correcting text and…