| Mobile Web

LLMs face limits in English; next battleground is language as multilingual design divides

Large language models' next competitive edge will depend less on scale and computing resources and more on multilingual design and reflecting local context, an assessment said. TechRadar reported that English-centric base models are showing limits as AI spreads globally and multilingual understanding becomes a key condition for sovereign AI. It said many models still translate English knowledge rather than reason within each language, raising risks for accuracy, adoption and trust in areas such as finance, healthcare and public services.