Advances in both Natural Language Processing (NLP) and Automatic Speech Recognition (ASR) have raised the question of whether AI-powered chatbots could be an alternative or supplement to flesh-and-blood human teachers in some situations. Can these tools really help learners acquire foreign languages?

From the malevolent Hal 9000 in Stanley Kubrick’s 2001: A Space Odyssey to the charming Samantha in Spike Jonze’s Her, computers that talk have shocked and seduced us in popular culture for many decades.

When Apple officially incorporated their voice assistant Siri into iOS in 2011, the reality of having an intelligent assistant that understood and obeyed our every word seemed one step closer for everyone.

The Amazon Echo “smart speaker” hit the market in 2014, and has been the dominant device in that field ever since. Other products in the same space include Google’s Nest, and Apple’s HomePod. Social networks also started to jump on the AI assistant bandwagon, with Facebook incorporating chatbots into its Messenger platform in 2016, and LINE releasing the Clova assistant in 2017.

Advances in both Natural Language Processing (NLP) and Automatic Speech Recognition (ASR) have raised the question of whether AI-powered chatbots could be an alternative or supplement to flesh-and-blood human teachers in some situations. Can these tools really help learners acquire foreign languages?

General purpose AI assistants for language learning

The applicable theory of language learning to bear in mind here is interactionism – the idea that languages are acquired by interacting with other speakers of those languages.

Even though smartphone and smart speaker based AI assistants haven’t usually been designed specifically with language learning in mind, some innovative teachers and researchers have used them for these purposes. One of the major issues to overcome here is the fact that these assistants aren’t optimized for non-native speech, and may struggle to correctly transcribe or understand it.

Research has shown that the tech behind these devices can recognize non-native speech to some extent (with Google Assistant recognizing 87% of learner utterances and Apple’s Siri recognizing 67% in one Japan-based study). 

So if a language learner is able to speak clearly enough for an AI assistant to understand them, what kinds of activities can be done to bring about further gains in language ability?

The applicable theory of language learning to bear in mind here is interactionism – the idea that languages are acquired by interacting with other speakers of those languages. 

Interacting with another person with language involves taking turns, negotiating meaning (figuring out what the other person is trying to say), and an information gap (transferring information from one speaker to another). There is no reason in theory why the tenets of interactionism cannot apply to human-computer interaction as well as human-human interaction.

General purpose AI assistants can stand in for human interlocutors in interview or quiz type activities, especially where the learner is asking the questions. However, the interaction tends to become one-sided, because AI assistants don’t ask questions unless programmed to do so. 

And while learners may receive implicit feedback on pronunciation or grammatical form where the AI doesn’t understand the question that has been uttered, they won’t receive explicit feedback unless they are using an app that has been specifically designed for language learners.

AI assistants specifically designed for language learning

There have been several attempts to develop AI assistants and other interactive speech apps specifically for language learners. Here we will take a look at some of these products and services, and evaluate their usefulness and effectiveness.

Duolingo Chatbots

Duolingo launched chatbots for its iOS app back in 2016, promising to help users “come up with things to say in real­-life situations”. Although the feature seemed to be well received by its users, it quietly slipped away and there is no sign of it returning yet.

Duolingo’s chatbots included a “Help me reply” feature, which would suggest words and phrases for the learner to use in their responses. The interactions with the chat bots would become more advanced as the users’ level progressed. There were some limitations to Duolingo’s chatbots though. For example, they only offered structured dialogues, as opposed to open-ended speech.

We hope to see a new version of these Chatbots from Duolingo as they have shown promising results for language learning outcomes.

Elai (ETS)

In December 2020, ETS released Elai (iOS/Android), an app that allows users to practice speaking about a range of topics and receive feedback on their speech.

Unlike Duolingo chatbots, Elai’s focus is on open speech. Users must respond to a prompt and record their answers within a 30 second time limit.

Elai offers a variety of feedback on learner speech, including the extent to which the learner repeated the same words; how often the learner paused in during their speech; and whether or not the learner used a lot of “filler” words, e.g. “uh”, “erm”, “ah”.

Elai attempts to improve the speaker’s vocabulary knowledge by providing a list of higher level words at the end of the exercise, which could also be used to respond to the prompt.

Elai is still in Beta status, and the extent to which it will be embraced by learners, teachers, and researchers is still an open question, but being developed by one of the world’s largest English language testing companies (ETS is behind the TOEFL and the TOEIC) certainly puts it in a strong position from the outset.

Buddy.ai

Buddy.ai (iOS/Android) focusses on the young English language learner market, and promises to “[provide] unlimited practice of spoken English.. to millions of students”.

The app offers a variety of language games and activities, including listen and repeat, question and answer, and interactive videos.

One of the drawbacks of the app, however, is that it only supports users with Russian, Spanish, Turkish and Polish as a first language. The app has a bilingual interface, and if the user has a first language other than one of these four, they will struggle to understand the instructions.

ELSA

Elsa (iOS/Android) is a mobile app that focuses specifically on improving the users pronunciation to help them “speak like an American” (although proponents of TEFL Equity might have something to say about this – should “American” be the target for all English learners?).

Through listen-and-repeat and interactive dialogue type exercises, Elsa teaches the user how words are blended together in casual speech, which in turn helps to improve the user’s fluency.

Summary

The principles of interactionism suggest that language learners can improve their skills simply by conversing with another speaker of the target language. However, there are issues to be overcome when using AI-powered virtual assistants for language learners, including lack of optimization for non-native speech, and lack of true discourse-level interaction.

Apps that specifically target language learners can do better when it comes to recognizing non-native speech, and offering more life-like interactions. 

English Central, for example, is one of the leaders in the recognition of non-native speech, and gives users instant feedback on their pronunciation and fluency while speaking lines from a library of thousands of videos.

However, many of the other apps discussed here focus on either niche segments of learners (e.g. Russian and Polish speaking children) or niche language language skills (such as fine-grain pronunciation problems). 

There is yet to emerge an artificially intelligent chatbot which can be used by all levels and all ages of learners that offers true human-like interaction and feedback.

In addition, student reactions to the recent COVID pandemic have shown that many students value face-to-face learning over online methods. Although chatbots and smart speakers could be a useful supplement to face-to-face or online learning with a human teacher, it seems unlikely that they will be a complete replacement for human teachers any time soon.

At this year’s CEGLOC virtual conference, I watched a couple of presentations about the role of Machine Translation (MT) in language teaching and learning.

They got me thinking again about a subject I’ve written about a few times before and also edited an article on.

Here are a few key assumptions about the intersection of MT and language teaching/learning:

#1 The accuracy/naturalness of MT is continuing to improve. The output produced by MT is approaching the point where it is virtually indistinguishable from the output produced by human translators.

#2 It is unacceptable for a student to rely solely on MT when submitting work for an assessed course of language learning. An example of this would be if a Japanese student wrote a report entirely in Japanese, pasted the report into an MT tool, copied the resulting English, and handed in the work in as their own (without even looking at the resulting English text)

#3 Notwithstanding #2, MT could have powerful pedagogical applications if used in the right way. 

#4 It is difficult, if not impossible, to completely prevent the use of MT without reverting to hand-written essays in exam conditions. Unlike plagiarism, MT cannot be easily detected by software. Although there are still some tell-tale idiosyncrasies of translations produced by MT (such as inappropriate grammatical subjects when translating Japanese to English, for example) such traits/mistakes are becoming less obvious as MT continues to improve

#5 The kind of behavior exemplified in #2 is clearly not a “CALL” technique. Computer Assisted Language Learning entails the use of computers to assist the learning of a language. Using MT to generate a piece of assessed work and handing it in sight-unseen is basically indistinguishable from plagiarism. But instead of passing off the work produced by another human intelligence as their own, the perpetrator is passing off the work generated by artificial intelligence as their own.

The above assumptions (if correct) raise some interesting questions, and force us to re-evaluate the reasons or motivations for learning a language.

The motivation for learning a language is often categorized into three main strands: integrative, intrinsic, and instrumental.

Integrative motivation compels language learners who wish to live in and integrate with a target language community. This kind of motivation might drive an American who wants to emigrate to and settle down in Japan, for example. 

Would such an individual be able to rely solely on MT tools and devices to achieve this goal? Could they whisper sweet nothings into their iPhone, and then place the iPhone on their pillow and allow it to translate and convey those sentiments to their significant other? Perhaps not.

Intrinsic motivation comes from inside the individual and often arises from a deep interest in the target language itself. Intrinsically motivated students are interested not only in the syntactic structure of the target language, but also in how speaking the language will change the way they perceive and interact with the world around them.

Could an intrinsically motivated individual leverage the power of MT to further increase their knowledge of the target language? I think so. Would they be happy to completely delegate to MT the task of translating their thoughts from L1 to L2? Would they want to miss out on the philosophical or cultural insights that learning another language can bring about? I think not.

Instrumentally motivated individuals simply treat the target language as a means to an end. They want to get a promotion or avoid being demoted. They want (or have) to do business with speakers of the target language. They want to quickly translate an email or subtitles for a video (e.g., rev.com). They want to pass an exam or entrance test for a particular business organization or academic institution. 

Could such an individual rely extensively on MT to achieve their aims? I think so. Would it be fair to allow them to do so, especially with regards to assumptions #2 and #5 above? Perhaps not. 

That question would need to be decided by the organizations and institutions involved, who are best placed to judge the skills and competencies they require from candidates.

Given all of the above, I tend to believe that the use of MT tools and devices will continue to increase, especially in situations where instrumental motivation is paramount, or time and money costs are significant. 

But in my role as a language teacher who has to assess the written and spoken output of language learners, there are difficult questions to answer with regard to the role that MT can or should play in the language learning process.

MT surely has many powerful pedagogical applications, but the temptation for time-pressed and sleep-starved students to rely solely on MT to produce the required output is high.

And then we’re into the familiar territory of plagiarism – passing off another’s work as your own. Something most academic institutions seriously frown upon.

So, those are my current thoughts on MT.

Would love to hear others.