Machine translation is almost a solved problem
But interpreting meanings, rather than just words and sentences, will be a daunting task
Vasco Pedro had always believed that, despite the rise of artificial intelligence (AI), getting machines to translate languages as well as professional translators do would always need a human in the loop. Then he saw the results of a competition run by his Lisbon-based startup, Unbabel, pitting its latest AI model against the company’s human translators. “I was like…no, we’re done,” he says. “Humans are done in translation.” Mr Pedro estimates that human labour currently accounts for around 95% of the global translation industry. In the next three years, he reckons, human involvement will drop to near zero.
Explore more
This article appeared in the Science & technology section of the print edition under the headline “The Babel wish”
Discover more
Humans and Neanderthals met often, but only one event matters
The mystery of exactly how people left Africa deepens
AI can bring back a person’s own voice
And it can generate sentences trained on their own writing
Carbon emissions from tourism are rising disproportionately fast
The industry is failing to make itself greener
Why China is building a Starlink system of its own
When it is finished, Qianfan could number 14,000 satellites, rivalling Elon Musk’s system
Lots of hunting. Not much gathering. The diet of early Americans
What they ate is given away by the isotopes in their bodies
Stimulating parts of the brain can help the paralysed to walk again
Implanted electrodes allowed one man to climb stairs unaided