Researchers at the National Bureau of Economic Research (NBER) have recently written a paper which examines the link between Machine Translation (MT) and an increase in international trade.
The paper, entitled “Does Machine Translation Affect International Trade? Evidence from a Large Digital Platform” reflects on the ability of AI (as one of the leading technological advances of our time) to impact trade, work and the economy, at an international level.
The study is the first major paper of its kind to scrutinise the material influence that AI and Machine Translation has upon trade and industry.
The NBER is a not-for-profit, private organisation is one of the most influential think tanks on economic policy in the US.
The paper, by Erik Brynjolfsson, a Professor at MIT’s Sloan School of Management, Xiang Hui, and Meng Liu was published in August 2018 and looks at how eBay’s machine translation engines have influenced international trade on the platform. The analysis used evidence “of direct causal links between AI adoption and economic activities.”
The study focusses on only the English to Spanish machine translation designed in-house by eBay developers, named eMT, which was introduced in 2014. The study considers the effect on international trade based on US exports.
Consumer Benefits from eMT
The introduction of eBay’s machine translation system can be seen to have significantly increased trade. Exports have increased on this platform by 17.5%.
Language specialist, Juan Rowda, who was a member of eBay MT staff, wrote an article in 2016, in which he discussed the feasibility of human translation for a platform such as eBay.
In this article he said that the company had in excess of 800 million articles listed, with an average of 300 words per item. His calculations suggested that it would take “1000 translators approximately 5 years to translate only 60 million listings that are available in Russia”.
For Rowda, the only possible answer to this problem was an advanced eMT.
eMT provides the right environment for products to be listed so that they can be classified, made cheaper, and offer options for advance title listings, which can all assist less experienced buyers in finding the right products.
Consumers also benefitted from eMT more than sellers, according to the study, due to the increase in language comprehension and a reduced price due to competition.
The paper also analysed the effect of eMT in Europe. British and Irish exports to France, Spain and Italy were found to confirm the earlier findings that eMT increases exports on eBay.
Trade and the Language Barrier
There have been other research papers that have investigated the link between language and trade, and these have also argued that trade is increased through language comprehension.
Overcoming this language barrier is high priority for many. In Europe, efforts to create a Single Market has led to the European Union investing millions in language technologies.
To this end, the Human Language Project, based at the German Research Centre for Artificial Intelligence (DFKI) is proposing a 10 year project (worth 1billion Euros), while another 5million euros have been allotted in grant funding for low resource language data and MT implementation.
As far as the researchers of the NEBR paper are concerned, MT has “made the world significantly smaller and more connected”.
MT and Increasing Trade
The paper also found that Machine Translation offered eBay a reduction in overall translation costs, which occurred due to increased accuracy, speed and translation quality, thus giving an improvement to listings and consumer match accuracy.
When compared to other applications for MT, such as self-driving cars, the NEBR researchers suggest that Machine Translation is far easier to implement, and this is why it has had such significant results with increasing trade on the platform.
The potential for ecommerce sites, such as eBay, to increase trade is also due to the sheer scale of projects undertaken by the company. For instance, the company generated more than $14billion of global trade in 2014 alone, in over 200 countries.
MT and AI solutions to language barriers are fast becoming an essential element for companies such as eBay, Amazon and so on, whose advances in these fields have attained various awards. The filtering down of these advances to other businesses, individuals and consumers is likely to increase international trade even more.
Recent Comments