AI in Language Learning: Today’s Uses, Benefits, and Concerns

Artificial intelligence, chatbots, large language models. These buzzwords have been popping up everywhere. Maybe you’ve dabbled with ChatGPT or you’ve asked an Alexa device a question. While it’s still new and constantly developing, artificial intelligence (AI) has the power to revolutionize many of our daily tasks.

Language learning is just one of the many areas where this type of technology is being used, and the potential it offers is mind-blowing. While the debate on exclusively using AI vs. a human teacher is still a hot topic, many language apps already employ AI to some extent. I’ve reviewed a lot of language apps, so I’m familiar with the selling points: AI voice recognition, algorithms that tailor the lessons to your level, and spaced repetition, amongst others. Well-known apps like Babbel and Duolingo have been using AI for years. 

But what is being developed now has the potential to blow all these traditional apps away. For example, students in Taiwan are practicing English with chatbots, eliminating the anxiety that new language learners feel when beginning to speak with others. And teachers have the ability to create customized lesson plans for a class or individual students in a matter of minutes. 

This all sounds amazing, but what’s the catch? As with any new technology, there are growing pains involved. Let’s take a look at how advancements in AI can help language learners and why some say it should be used with caution.

What is Artificial Intelligence?

AI enables computers and machines to perform some tasks that typically required human intelligence in the past. It involves the development of computer programs that can learn from data, adapt to new information, and make decisions or perform actions without explicit human programming. AI systems use different techniques, such as machine learning, neural networks, and natural language processing, to understand complex information, recognize patterns, and solve problems.

AI works by inputting large amounts of training data, which is then analyzed by the program for correlations and patterns. These patterns are then used to predict responses to future prompts. Therefore, the quantity and quality of the training data matter because that is what the output will be based on.

To put it simply, large amounts of data go in, get analyzed and processed, and answers come out based on what the machine has “learned” from that data.

How We Already Use Artificial Intelligence for Language Learning

Artificial intelligence is already widely used in the language-learning industry. From language apps to chatbots, there’s a range of applications.

A survey of close to 6,000 language learners in Sweden showed how popular these technologies are and whether students thought they were helpful for learning a language. It compared already-established AI language tools with up-and-coming chatbot technology. Take a look below to see the survey results.

One of the most common applications of AI is automatic language translators or machine translation. Some popular examples include Google Translate, DeepL, and Reverso. They use algorithms to translate a text from its source language to a different one. They’re helpful for translating basic words and phrases, but the accuracy of the translation can vary widely.

Other uses include writing correction tools, such as Grammarly, and speech-to-text transcription. While both can help native speakers and language learners alike, they have particular benefits for the latter. Speech-to-text recognition converts spoken language into written text while speech recognition programs analyze a speaker’s pronunciation and give feedback on the accuracy of the pronunciation. In language learning, it’s perhaps most beneficial in assessing pronunciation.

However, I’ve saved the best for last. If you’re using an app to learn a new language (which, if you’re a Langoly reader, you probably are), it’s almost certainly AI-powered to some degree. They are utilizing AI to offer personalized learning experiences by analyzing the strengths and weaknesses of each learner to provide customized lessons and practice exercises. They can also track a learner’s progress over time and adjust the level to their needs.

For one example of how AI is being used to enhance the language learning experience in apps, check out the MosaLingua video below:

But MosaLingua isn’t even close to being the only one. Duolingo was recently named by TIME Magazine as one of the 100 most influential companies. They praised the app for the incorporation of its own AI model, called Birdbrain, and for adding newer AI tools like GPT-4, which are “allowing users scenario-based ways to practice, like going furniture shopping, asking a friend to go for a hike or ordering coffee at a café in Paris.” This helps bridge the gap between learning a language from an app and living in a country where the target language is spoken.

Chatbots, Natural Language Processing (NLP), and What We’ll See in the Future

According to IBM, natural language processing (NLP) refers to “the branch of computer science—and more specifically, the branch of artificial intelligence or AI—concerned with giving computers the ability to understand text and spoken words in much the same way human beings can.” NLP is most commonly seen in chatbot features, such as the ones you use on your phone. It’s the reason why you don’t have to give Siri an exact command but can talk to her like she’s a real person.

Imagine that you’re just starting to learn a language. We all know that speaking is one of the most important skills, but also one of the hardest to develop. If you have the ability to speak to an NLP program and receive feedback and corrections without any judgment or frustration from a human, it seems like a perfect way to practice (in theory).

That’s why we’re seeing new platforms pop up that deal specifically with AI chatbots for language practice. These are services that goes a step beyond run-of-the-mill language-learning apps because they’re built specifically for language output. One example is GoPenPal. You can message the chatbot and receive instant corrections and translations. It also breaks down exactly what the chatbot said so you can understand the messages. 

These are only the beginnings of what we may see in the future of AI in language learning. Imagine an overcrowded classroom where one teacher has to correct 30+ students. While a chatbot isn’t perfect, it can give personalized corrections to a language student, especially at the beginner level. It can also help learners get over the common fear of speaking a new language. Chatbots don’t feel emotion and give corrections based on their training data.

AI Sounds Great! But What’s the Catch?

The benefits of using AI for language learning are numerous. But as with any new technology, there are some concerns. The most important, as any AI developer will tell you, is that this technology cannot replace an actual person. One of the biggest complaints about AI is a phenomenon called “hallucinations.” This means that the AI model gives you an incorrect answer. 

However, if you’ve received a series of correct answers and then receive an incorrect one, how will you know that it’s incorrect? The technology seems very confident in its response, so unless you do some investigation, you won’t know that you received an incorrect answer.

Another drawback of chatbots for language learning is that they have trouble understanding the finer points of natural language. This means they might not always get what the user is saying, and their responses could be wrong or confusing. As a result, learners may end up learning incorrect language usage without even realizing it. Since chatbots can’t pick up on context or subtle language differences, they might not provide the level of understanding needed to fully grasp the complexity of conversational language.

But there are deeper issues that need to be addressed too, such as data protection and lack of regulation. Italy became the first country to ban ChatGPT because of concerns that it violated data protection laws (although this ban has since ended). Even OpenAI chief executive Sam Altman testified before the US Congress that artificial intelligence could “cause significant harm to the world.” However, he went on to promise that OpenAI will continue to release the technology as planned.

Final Thoughts: AI in Language Learning

As with any new technology, there are both excitement and skepticism. However, it seems that AI technology is here to stay, so it’s better to embrace it rather than fight it. It’s important to realize both its benefits and also the problems and limitations it causes. Using AI well is a skill, and it takes training and experience to learn its best uses. 

Most language apps and other language platforms that we’re using have already incorporated this technology into their platforms to some extent. It seems like the choice to use it (or not) has already been decided for us. And its uses will most likely continue to expand.