Sydney: (The Conversation): From real-time speech translation in video calls to auto-dubbing on social media platforms, artificial intelligence (AI) is rapidly removing language barriers and making communication across languages nearly instant.

Tools developed by companies such as OpenAI, Google and Meta now offer near-real-time translation across dozens of languages, raising a fundamental question: if machines can translate faster and more accurately than humans, is it still worth spending years learning another language?

The argument for AI is compelling. Humans have historically delegated cognitive tasks to tools—from writing systems reducing memory load to calculators simplifying arithmetic. AI, in this sense, is a continuation of that evolution, offering powerful assistance and accessibility.

However, researchers argue there is a key distinction between using tools to enhance learning and relying on them to bypass it entirely.

The value of effort in learning
Psychologists describe “desirable difficulties” as challenges that may feel inefficient in the short term but lead to stronger long-term learning and retention. Struggling with grammar, vocabulary and sentence formation engages brain systems linked to memory, attention and cognitive flexibility.

Such sustained mental engagement is also associated with cognitive resilience—the brain's ability to maintain function with age. Managing multiple languages strengthens this by requiring the brain to constantly switch, monitor context and resolve competing meanings.

Experts caution that these benefits are harder to achieve when learners rely entirely on translation tools.

What research on multilingualism shows
Studies on multilingualism offer mixed evidence. While some suggest a “bilingual advantage” in attention and working memory, others find limited or no overall cognitive difference.

A recent study of 94 adults aged 18 to 83 found that multilingual individuals performed similarly to monolinguals in most cognitive tasks. However, those with richer multilingual experience showed improved visuospatial working memory, particularly among older adults.

This suggests that multilingualism does not broadly enhance all cognitive functions but may help preserve specific abilities over time. Other research also links multilingualism with delayed onset of Alzheimer's disease and healthier cognitive ageing, though mechanisms remain under study.

What AI translation cannot replace
AI translation is highly effective for speed and accessibility but operates through pattern recognition rather than lived understanding. It often struggles with cultural nuance, humour, tone and context, especially in less-represented languages.

Researchers argue that language learning is not just about communication, but about cultural participation. It involves understanding values, identity and context that cannot be fully captured through automated translation.

As one participant in the study noted, different languages can shape different emotional and cognitive experiences—something AI tools cannot replicate.

Beyond translation
While AI will continue to transform language learning through personalised tools and instant feedback, experts say it cannot replace the deeper cognitive and cultural engagement involved in learning a language.

The researchers conclude that language learning remains important not just for communication, but for how people think, express themselves and understand the world.