How AI translation could unseat English as the lingua franca of the business world

Anyone who has traveled to a country where the language spoken is not their native one knows that not conversing fluently (or at all) can turn even a VIP into a second-class citizen. Einstein himself would have struggled to express his intelligence in, say, Farsi. In one of my favorite episodes of Modern Family, Sofia Vergara’s character Gloria says in frustration, “You don’t know how smart I am in Spanish!” Even fluent speakers can face bias if they have an accent because of certain underlying perceptions that your language skills are correlated with your intelligence.

No one deserves to feel like a second-class citizen, and English as the common lingua franca just doesn’t work for everyone. Luckily, advances in AI-powered translation can help us overcome language inequities around the globe by helping everyone access information in their native language. It isn’t just good for people. It’s good for business.

Language inequality is bad for business

In an English-centric world, achieving global diversity and inclusion is a big challenge for business, and language equality plays a huge part.

What is language equality? In theory, it’s people’s ability to access info, products, or services equally, independent of language. In practice, it’s making that ability as strong in “developing” countries as in “developed” nations.

Or in developed nations that are less wealthy than their closest neighbors, like my native Portugal. Because of the country’s modest economic size, compared to most of Western Europe, many online companies have limited (or no) presence in Portuguese. British Airways, for instance, only offers customer service in Portuguese on weekdays during business hours—and they’re a global airline with enormous operations in Europe. What’s more, there are almost 230 million native speakers of Portuguese worldwide, the vast majority of them in Brazil (where, yes, British Airways also offers flights). It’s the sixth most spoken language in the world.

As a test, our company sent an email in Portuguese to the customer service departments of several companies with a simple question: “How do I change my password?” Some of them responded as though we’d written gibberish. The majority said, “We only provide customer service in the following languages,” and Portuguese wasn’t one of them. Companies typically set up their support operations in the languages with the highest contact volume, and unless you happen to come across a native-speaking agent, you’re out of luck.

What is the effect of language inequality for business? Missed opportunities, frustrated customers (or potential customers), and, ultimately, lower profits. A 2020 study from Common Sense Advisory showed that 40% of customers will not buy products from companies that don’t support their language. Three-quarters of buyers said they want product information in their native language. Prior to smart, AI-driven translation, it was complicated and costly to have a robust customer service operation in multiple languages. Now there’s just no excuse. If you’re offering a 24-7 support center, those agents can handle many languages thanks to fast, increasingly accurate translation software technology. Using AI, an English-speaking customer service agent can reply to an email or chat to a Portuguese customer, in Portuguese.

Other language inequalities

Education, tourism, and coding are other areas that suffer from language inequality. Although some attempts to remedy language inequality exist in early primary education (particularly in Europe) only a few languages (and mostly English) dominate in the online sphere. At a time when global cooperation is crucial to our very survival, most large, free online courses (or MOOCs) offered by prominent universities are taught in English, and the “thought leaders” teaching them are primarily from the U.S., the U.K., Germany, or the Netherlands. The current pandemic has also thrown this educational inequality into sharp relief—and fueled a huge demand for subtitles on digital content in English.

In other situations, the English hegemony is turned upside down. Want to know the best restaurant in Lisbon? Locals will be happy to tell you, in Portuguese. The best resources for travelers are usually written by locals, in their native tongue. Beyond tourism, truly grasping what’s happening in another country requires an understanding of local news. A nuanced understanding politically, economically, and culturally demands the ability to digest incoming information in a different language. How can people ever hope to gain knowledge about, say, Japan, if everything they read or listen to is created by non-Japanese? Fortunately, technology can enable local information—from news reports to restaurant reviews—to be translated into virtually any language. We need to reach a tipping point where this technology is used systematically on a global scale.

Software is another great example of how English-language hegemony hurts innovation, particularly in technology. StackOverflow is hands-down the biggest resource in the world for software developers. The vast majority of these developers participate in the English-language version of the site. While StackOverflow has localized sites (such as StackOverflow Brazil) the developer communities are limited to local knowledge only, and can’t access the benefits of learning from developers across languages. That leaves a lot of cross-cultural collaboration off the table.

The economic power of some countries ensure language and cultural dominance. That’s not new. But this approach stifles innovation and prevents the spread of knowledge around the world. For example, when Alibaba required that their devs learn English because all the tools and coding languages were English-based, the Chinese started building their own tools, in Mandarin, for Chinese people. And where does that leave all the devs who don’t speak Mandarin but would like to access these tools? Locked outside the circle of knowledge—unless they start learning the language. There has got to be a better way to accelerate innovation that doesn’t involve building more linguistic silos.

How can we get to a point where we’re language independent, and the tools we build are language agnostic? Even the brightest minds can rarely achieve a high level of fluency in more than two or three languages. While technology (namely the internet) has exacerbated language inequality, it can also hold the key, through rapid translation assisted by artificial intelligence. The last year alone saw major breakthroughs in AI-centered language models, from Facebook’s M2M-100 translation model to Google’s MT5 and OpenAI’s GPT-3. Eventually, we’ll see a common multilingual machine translation model. It’s time to start putting its intelligence to use.

Vasco Pedro is CEO and Co-Founder of Unbabel.