Emerging Technologies

Machine learning is tearing down language barriers. What does this mean for trade?

Improved machine translation means people who speak English will soon find themselves in direct competition with talented people who don’t. Image: REUTERS

Richard Baldwin
Professor of International Economics, Graduate Institute, Geneva
Our Impact
What's the World Economic Forum doing to accelerate action on Emerging Technologies?
The Big Picture
Explore and monitor how The Digital Economy is affecting economies, industries and global issues
A hand holding a looking glass by a lake
Crowdsource Innovation
Get involved with our crowdsourced digital platform to deliver impact at scale
Stay up to date:

The Digital Economy

This article is part of: Annual Meeting of the New Champions

We all know what “intelligence” means and what “artificial” means; yet, put them together into “artificial intelligence” and we get confusion, trepidation or maybe even laughter. The phrase “artificial intelligence” may not ring the same bells for everyone. One form of artificial intelligence that should be ringing everyone’s bells is called machine learning. Machine learning is transforming our world faster than most realize; in particular, it is transforming trade.

Machine learning represents a radical change in the way computers “think”. Before machine learning, we taught computers to do things with computer programs that explained, step-by-step, what the computer should do. This limited computers to mimicking human thinking only where we understand step-by-step how humans think. For example, we understand how we do arithmetic and algebra since this is conscious thinking. By contrast, we have no idea how we recognize a face or keep our balance when running over an open field since this is unconscious thinking. Before machine learning, computers could only perform conscious-thinking tasks since those were the only ones we knew how to program. Unconscious-thinking tasks were beyond the reach of computers since they could not be programmed.

Machine learning changed this by skipping the step-by-step programing. With machine learning, the computer (the “machine” part) uses a very large amount of data to work out how to best guess the solution to a particular problem (the “learning” part). Thanks to exponential advances in computing, power and access to preposterous amounts of data, computers trained by machine learning routinely achieve human-level performance on unconscious-thinking tasks – like recognizing handwriting, speech or bone breaks in X-ray images. One game-changing application of machine learning is machine translation.

What is machine translation?

Machine translation – machine learning applied to language – has gotten amazingly good in recent years. And it’s free, works instantly and is getting better every day. This one development will change the way the modern world works, since language barriers have been an important hindrance to international commerce since time immemorial.

Machine translation is not some exotic future technology in its beta testing phase. It is already on your smartphone, laptop and tablet. Free apps like Google Translate and iTranslate Voice now work quite well between major languages. Other smartphone apps include SayHi and WayGo. And it is widely used. Take Google, for instance: it does a billion translations a day for online users. Microsoft introduced automatic, instant translation into Outlook email; Twitter offers translation on most foreign-language tweets. YouTube has instant machine translation for many foreign-language videos: just go to the settings “gear”, click on “captions”, and choose “auto-translate”. Instant, free, spoken translation is also possible on Skype: the add-on Skype Translator will allow you to understand foreign-language speakers you are Skyping with. It is not perfect, but being able to talk freely with someone who’s speaking a different language is nothing short of marvellous. And it will be a massive stimulus to trade.

Language’s impact on trade

Economists’ mainstay approach to measuring the impact on trade flows of things like a common language and distance is called the “gravity model”. It predicts that, like the force of gravity, bilateral trade flows are boosted by the economic mass of the selling nation and the economic mass of the buying nation, but diminished by the geographic distance between them. In empirical studies, economists find that language barriers dampen trade in a spectacular manner, while estimates suggest that sharing a common language boosts trade among nations by about 50%. This may sound high to many, but probably rings true to most international business leaders who find on a daily basis an almost unimaginable range of problems arising when buyer and seller can’t talk directly.

The implications are clear. As machine learning tears down the language barrier between major languages, world trade flows should rise – a lot. Since machine translation is getting so good so fast, the impact should already become noticeable within the next few years.

The impact has already been documented in one particular form of trade: online international freelancing, or what I call “telemigration” in my forthcoming book, The Globotics Upheaval. That is, people sitting in one nation but working in another nation. It is online freelancing done international and is mostly coordinated via matchmaking platforms that are like eBay, but for services rather than goods. People who want to hire freelancers and people who want to freelance register on these sites, and the platform helps them meet, while simultaneously making it easier to pay, manage, and communicate with each other. Although there are millions of freelancers registered, since most of the work is in English, the ability to speak reasonable English is a big barrier for non-English speakers to this new form of globalization.

Of the 7.2 billion humans on this planet, about 400 million speak English as their first language. Adding in a generous estimate of non-native English speakers brings the number up to about a billion English speakers. With machine translation becoming so usable, people who speak English will soon find themselves in much more direct competition with talented people who don’t. The result will be a talent tsunami. With machine translation, a vast number of people will be speaking English, or other rich-nation languages like French, German, Japanese, or Spanish, well enough to take part in this job market.

Just think about China. Since around 2001, China has been producing over 8 million graduates per year – more than twice the number graduating in the US. These graduates find work, but it is often in part-time or low-paid jobs for which a university degree is not necessary. Just imagine the increase in competition that will take place now that they can (via machine translation) speak good enough English and sell their brain power over the internet to the US, Europe, Japan and other rich nations.

One recent academic study by Erik Brynjolfsson (of The Second Machine Age fame) estimates that the introduction of machine translation has already increased international freelancing by over 17%. As more people and companies adapt to the new possibilities, the impact could be much, much larger.

Why now?

Machine translation used to be a joke. Just two years ago, it was little more than a party trick for bilingual people. Now it is rivalling average human translation for popular language pairs: Google research used humans to score machine translations on a scale from zero (complete nonsense) to six (perfect). In 2015, Google Translate got a grade of 3.6 – far worse than the average human translator who got scores around 5.1. After a massive upgrade that came in 2016, Google Translate now hits numbers like 5. The capabilities of machine translation are advancing by leaps and bounds.

So why has machine translation suddenly become a real force to be reckoned with? The simple answer is: data. In 2016, the UN posted online a dataset with nearly 800,000 documents that had been manually translated into the six official UN languages: Arabic, English, Spanish, French, Russian, and Chinese. The same year, the EU released a series of official translations into its many official languages, and the Canadian parliament released its debates that had been hand-translated between English and French. All this data went into training machine-translation algorithms. The result? Explosive progress.

Companies and governments should take note. One of the greatest barriers to human cooperation is falling at an explosive pace. As usual, the changes will create pains and gains. It is time for us all to start taking steps to profit from the opportunities and address the challenges.

Have you read?
Don't miss any update on this topic

Create a free account and access your personalized content collection with our latest publications and analyses.

Sign up for free

License and Republishing

World Economic Forum articles may be republished in accordance with the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International Public License, and in accordance with our Terms of Use.

The views expressed in this article are those of the author alone and not the World Economic Forum.

Related topics:
Emerging TechnologiesJobs and the Future of WorkEconomic GrowthTrade and Investment
World Economic Forum logo
Global Agenda

The Agenda Weekly

A weekly update of the most important issues driving the global agenda

Subscribe today

You can unsubscribe at any time using the link in our emails. For more details, review our privacy policy.

What is the 'perverse customer journey' and how can it tackle the misuse of generative AI?

Henry Ajder

July 19, 2024

About Us



Partners & Members

  • Sign in
  • Join Us

Language Editions

Privacy Policy & Terms of Service

© 2024 World Economic Forum