Is Language Learning Still Needed?
Machine Translation has Greatly Improved with AI Technology
AI technology has largely facilitated the development and improvement of machine translations. For example, Google has been developing translation methods using artificial intelligence (AI) technology and achieved great success. You may have experienced AI-powered machine translation if you’ve ever tried Google Translate, –Google’s end-to-end deep learning machine translation system, known as neural machine translation. It learns from a large amount of data over time about how to simulate human translations, which means producing languages sounding more like real human language (VOA, 2016). The accuracy and time-efficiency of translation has been greatly improved. Therefore, this seems to pave a way for people to get rid of the need of learning foreign language on their own.
However, there are still many voices saying that machine learning can never replace language learning. On the one hand, leaning a language is never equal to translation. Learning a new language is about learning to communicate with others in that language. This requires that learners must also have knowledge of cultural contexts and thinking in that language in order to comprehend real communication (Haugh & Head of School of Languages, 2018). Besides, learning a language is an attractive learning activity which has its intrinsic characteristics that motivates learners. In the process of learning a new language, there’s a sense of accomplishment learners can get, leading users to build confidence and overcome challenges themselves (Oroujlou & Vahedi, 2011). Thus we see social functions have been implemented in many language learning tools. For instance, many people join language learning communities on the learning platform to increase their social circles.
Here is the video of renowned translator David Bellos discussing What Lost in Machine Translation:
This is YiWang, Liulishuo’s founder and CEO, talking about whether AI will replace English teachers:
AI Empowers Language Learning Tools
In addition to facilitating machine translation, AI plays another important role in language learning. That is, offering AI tutor services to language learners who have limited access to appropriate language environment or instructor resources. Here I am going to introduce two most representative AI-powered language learning products–Liulishuo and Edwin. These two products represent the current mainstream presence of AI in language learning tools. Liulishuo is the most successful AI-powered products in Chinese English-learning market and it has went on Nasdaq stock market in 2018. Edwin is a multiple language learning product and part of Google Google Assisstant ecosystem.
Liulishuo-Your personal AI English Tutor
Liulishuo (also known as LingoChamp)–“speaking fluently” in English, has a featuring AI teacher service that utilizes cutting-edge deep learning and adaptive learning technologies, big data, as well as well-established education pedagogies, such as gamified learning process, peer interactions, and etc to deliver a user-centric, personalized and effective English learning experience accessible to anyone, anywhere, at anytime.
How does Liulishuo use AI?
Based on Liulishuo’s world’s largest database of Chinese people speaking English that covers “a broad range of geographic distribution and proficiency levels(2018)”, they created a Chinese-speaking-English recognition engine with the highest accuracy and an evaluation engine which provides users with ratings and feedback. “As of mid-2018, the company had recorded approximately 1.3 billion minutes of conversation and 17.5 billion sentences. To put that into perspective, it would take you 2,473 years of constant listening to get through all that banter (2018).”
Liulishuo’s AI teacher is not human, it is a system that relies on user interactions to make decisions. Through deep learning and other AI technologies, it selects only relevant content from a huge library and recommends it based on the student’s level (2018).
For example, they have a special function tackling the IELTS exams which is comprised of pronunciation, grammar, vocabulary, and fluency tests. Liulishuo recommends individualized suggestions of courses to users according to their scores in the four areas. “The scoring algorithms have passed the Turing Test. The variation in score between the AI tutor and a human examiner is lower than between two human examiners (2018)”.
“ETS TOEFL testing has proven Liulishuo AI teacher can improve learning efficiency by three times. For example, if it used to require 100 hours of learning to reach a certain level in the CEFR standard, we would only need about 36 (2018).”
Edwin-AI-powered English Tutor Backed up by Google
Edwin is an AI-powered English language learning course that functions through Facebook messenger. According to the company, it has 800,000 registered users, of whom the majority are from Mexico. It offers individual users with a personalized program that includes both an AI tutor and professional English tutors to help prepare them for standardized English on a foreign language test, such as TOFFEL(Nazerian, 2018). Having combined machine learning, natural language processing, and speech technology(HRT News Desk), what Edwin is aiming at is to deliver personalized and adaptive learning plans to global English learners at a reasonable price. Unlike Liulishuo that targets on Chinese learners only, Edwin supports learners with various mother tongues, such as Spanish, Japanese, Korean, Russian, Hindi and Arabic (Nazerian, 2018).
Google is investing Edwin, incorporating it into its Google Assistant ecosystem. That means Edwin will have access to consult from Google engineers, and get marketing resources from the most powerful tech company int he world(Staff, 2018). Edwin is going one step further than chatbot by which students record their words and check pronunciation. With Edwin, learners would be able to speak directly with the AI assistant to practice pronouncing words. Google Assistant will also let students practice conversational dialogue, which is considered “impractical” with a chatbot at current stage(Nazerian, 2018).
Focus Question：
- Will AI and machine learning powered translation machine replace the need of learning a foreign language?
- What is your opinion about the current AI-powered language learning tools? Do you think it is just a selling point? Or do you think AI would be the game-changer for language learning?