San FranciscoOpenAI paused training over safety concerns last month.
Its safety tests run mostly in English, leaving other languages effectively unchecked.
It was the first time a major AI lab voluntarily froze its own model, the outlet reports.
It’s unclear what triggered the pause, or when training will resume.
How each outlet framed it
- Rest of World
- AI safety gaps concentrate in Silicon Valley; multilingual chatbots fail health queries in Africa/Asia, affecting diagnoses and treatment
Sources: Rest of World