In this blog series, resident IEX Senior Consultant, Peter Debenham, shares his predictions for how certain technologies will shape our economic and social environment in the future. This week: Artificial Intelligence and Natural Language Speech Processing.
Artificial Intelligence (A.I.) is one of those terms which seem to be constantly mentioned but are rarely explained. At one extreme, people think of intelligent robots which may be able to do all we humans can, and more; whether or not this is a utopian or dystopian vision depends on who you speak to.
At the other extreme, what we certainly would have called ‘A.I.’ fifty years ago is all around us now. One example where A.I. is used daily is Amazon’s transactional A.I. which predicts what we are interested in purchasing from our online behaviour, driving its near dominance of online shopping.
Other examples of applied A.I. are the diagnostic analysis of radiological scans, especially in oncology and cardiology where computers are better at reading large numbers of images than human radiologists, or the ability of TomTom and Google Map's to predict journey times based upon large amounts of historical trip data. We are all getting used to machine intelligence being part of our daily interaction with computing devices.
What may be more interesting is what happens when instead of A.I. being just involved in the human to machine interface it starts mediating in normal day-to-day human to human communication.
More International Trade and less English
In the not too distant future, speech recognition and translation will become sufficiently good to allow near normal communication in a language you do not know via ear-buds. Those who only learn languages “because they must” will stop doing so (bye-bye language teachers). Probably more interesting "bye-bye English," as a second language (after all, why does everyone in the West learn English? – it gets the maximum communication for the minimum learning buck).
Balanced against this is a likely great uptick in international trade (and self-guided tourism) as language itself ceases to be a barrier. A.I. can also lend itself useful here.
Currently, many people struggle to learn second, third, and more languages because otherwise when they travel they are unable to communicate with those who surround them. Efficiency means that for a great many people they pick to learn one of a relatively few languages.
English is only the third most commonly spoken native language with 380 million native speakers with, Mandarin Chinese having nearly 920 million speakers and Spanish 460 million. But English is the great winner in the second speaker stakes. It is the first language for around 380 million people in the world but the second language of 750 million more, giving it over 1.1 billion speakers. By comparison, the next most common second languages, Hindi, and Arabic, only have 274 million. French is fourth here with just over 202 million.
Now imagine rather than everyone just having a mobile phone we all have earbuds and microphones coupled to A.I. systems capable of understanding natural language and offering an instant translation of what is being said. It is not hard to imagine the technology: low-power IoT devices with high-speed 5G bandwidth providing the interface.
Siri, Alexa and Google Assistant already show natural language speech recognition. Google Translate and Skype translator can already offer immediate translation which with the power of machine learning is continuously improving.
But if imagining the technology is easy, imagining the effect is much harder. How many people view learning a second language as something they want to do for fun?
If second languages are no longer required, will they still be learnt by any except a minority of people? This might be answered by observing that over 3/5ths of Britons only speak English, whilst in the EU as a whole almost 3/5ths speak at least one foreign language, a number pulled down by the UK’s poor foreign language skills. Britons do not learn a foreign language because English is already the best language to use if you need to communicate outside the United Kingdom; something, not the case if you are from (say) Finland (where over 90% speak a foreign language).
A vast number of resources and many people’s employment are dedicated to learning languages and providing professional translation services. Currently, it is estimated that over 250,000 native English speakers work as English teachers abroad, not to mention many more non-native English speakers working as language teachers. Once a seamless immediate translation is possible will we as a society decide to dedicate the time and money elsewhere instead?
Maybe getting that TEFL qualification will not provide a job for life after all?