Despite the worries of some language teaching professionals that advances in AI will completely negate the need to learn foreign languages, this seems unlikely to happen in the near future. In this short essay, I will examine what is required to be an effective communicator in a foreign language and consider whether AI will take the mantle from humans in this endeavour.
In their seminal paper, Canale and Swain (1980) list four core competencies required to use and understand a foreign language:
grammatical competence: knowledge of the linguistic code of the target language, including vocabulary, morphology, syntax, semantics, and phonology;
sociolinguistic competence: knowledge of the social rules of language use, including cultural norms, appropriateness in terms of topic, audience, setting, and the relationship between participants;
discourse competence: knowledge of how sentences in discourse are connected to form a cohesive and coherent whole;
strategic competence: knowledge of communication strategies that can be employed to overcome difficulties in communication, including paraphrasing, circumlocution, and gestures.
Using the paid version of ChatGPT in voice mode (OpenAI, 2023), you can now have unnervingly realistic conversations with AI, which include “ums” and “ahhs,” references to previous topics, cohesive mechanisms, paraphrasing, changes in topic, varying levels of formality and politeness, follow-up questions, and strategies that encompass almost the full spectrum of Canale and Swain’s (1980) communicative competencies. Moreover, ChatGPT can do this in phonologically accurate voices that span over 50 languages, seamlessly code-switching as it goes. This may be jaw-dropping, but it remains doubtful whether technologies like ChatGPT will ever completely negate the need to learn foreign languages, even when such technologies are embedded in the devices we carry with us everywhere.
Universal translator devices have long been imagined by sci-fi writers, from Star Trek’s Gene Roddenberry to The Hitchhiker’s Guide to the Galaxy’s Douglas Adams. While a true universal translator that instantly and flawlessly translates any language is yet to be invented, pocket translators like the Pocketalk (Figure 1) or the Timekettle X1 AI Interpreter Hub (Figure 2) are now a reality. These devices allow the user to speak into them and have the onboard AI-powered software translate their utterances into the first language of their interlocutor. Devices like these are increasingly visible in countries such as Japan. For instance, the university where I work is trialling them at the IT support desk, and some Japan Rail (JR) staff have been equipped with them. Kirkpatrick (2020) notes that “these devices represent a significant increase in accuracy and functionality above manual, text-based translation applications such as Google Translate” (p. 15).
Figure 1
The Pocketalk Translator
Figure 2
The Timekettle X1 A1 Interpreter Hub
In addition to carrying such devices, we may soon be wearing them. A few years ago, Google teased us with a video showcasing its new Project Iris smart glasses (see Figure 3), which use augmented reality (AR) and machine translation (MT) to translate and render subtitles for spoken foreign languages right in front of the user’s eyes (Google, 2022). Although these glasses eventually failed to materialise (Hollister, 2023) it seems inevitable that devices like these will continue to be developed, and eventually become as ubiquitous as smartphones are today. However, would we really want to rely on them to mediate our multilingual relationships?
Figure 3
Google’s Project Iris Augmented Reality Smart Glasses
I recently spoke with a 30-something Japanese female friend who was lamenting her last failed relationship with a foreigner because “I can’t speak English well, and he couldn’t speak Japanese at all,” adding that “the relationship failed because we had to rely on technology to communicate.” Pocket translation devices can be intrusive and clunky, and whispering sweet nothings into one and having it translate and robotically recite your sentiments to your significant other could certainly be a barrier to romance. However, the voices are becoming less robotic (e.g., voices developed by ElevenLabs or Synthesia), and the devices are becoming less clunky and intrusive. In situations that do not normally involve the intricate nuances of romantic relationships, such as asking a member of JR staff which is the correct platform for Shinjuku, pocket translators could work just fine.
Therefore, will advances in AI negate the need to learn foreign languages? Almost certainly not. There are still plenty of reasons to actually learn a foreign language (where learn means to be able to use and understand it in a variety of situations in an unassisted way). We regale our students with these reasons in introductory English classes, often stating: “You’ll broaden your horizons,” “You’ll bolster your employability,” “You’ll think differently,” “You’ll make new friends and influence people,” or “You’ll see the world in all its Sapir-Whorfian glory” (Sapir, 1929; Carroll, 1956). Those with strong enough instrumental, integrative, or intrinsic motivation (Gardner & Lambert, 1972; Noels et al., 2000) will always be compelled to learn foreign languages, especially a foreign language such as English, which serves as the de facto global lingua franca (Crystal, 2003; Jenkins, 2006). Also, let’s face it, being able to speak a foreign language is much cooler than relying on technology to translate for us. It is also, apparently, more romantic and almost certainly necessary to maintain any non-superficial human relationship.
Translation devices will, however, diminish the urgency with which foreign languages are learned due to the reduced risk of communication breakdowns (granted, they may introduce other breakdowns). These devices will be sufficient in a variety of situations—both business and personal—where multilingual communication is essential. However, they will likely never match the capabilities of human interpreters, translators, or speakers of foreign languages. Humans have a wide range of expressive and communicative techniques, alongside cultural, situational, and emotional understanding, that a unidimensional translation device cannot hope to match. That is, at least not until AI is fully embodied and conscious, which remains firmly in the realm of science fiction.
Addendum
On May 13th, 2024, a few days after this article was submitted for copy editing, OpenAI released GPT-4o, which they describe as “a step towards much more natural human-computer interaction—it accepts as input any combination of text, audio, and image and generates any combination of text, audio, and image outputs” (OpenAI, 2024, para. 1). Although I believe that the assertions in this article—that advances in AI will not negate foreign language learning—still hold, this release shows just how quickly AI technologies are developing. By the time this article is published, there will likely be even more advanced models, with even more human-like capabilities.
References
Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics, 1(1), 1–47. https://www.uefap.com/tefsp/bibliog/canale_swain.pdf
Carroll, J. B. (Ed.). (1956). Language, thought, and reality: Selected writings of Benjamin Lee Whorf (2nd ed.). MIT Press.
Crystal, D. (2003). English as a global language (2nd ed.). Cambridge University Press.
Gardner, R. C., & Lambert, W. E. (1972). Attitudes and motivation in second language learning. Newbury House.
Google. (2022, May 12). Breaking down language barriers with augmented reality [Video]. YouTube. https://www.youtube.com/watch?v=lj0bFX9HXeE&t=28s
Hollister, S. (2023, June 28). Google has reportedly killed its Project Iris augmented reality glasses. The Verge. https://www.theverge.com/2023/6/27/23776144/google-project-iris-ar-glass...
Jenkins, J. (2006). Current perspectives on teaching world Englishes and English as a lingua franca. TESOL Quarterly, 40(1), 157–181. https://doi.org/10.2307/40264515
Kirkpatrick, K. (2020). Across the language barrier. Communications of the ACM, 63(3), 15–17. http://dx.doi.org/10.1145/3379495
Noels, K. A., Pelletier, L. G., Clément, R., & Vallerand, R. J. (2000). Why are you learning a second language? Motivational orientations and self-determination theory. Language Learning, 50(1), 57–85. https://doi.org/10.1111/0023-8333.00111
OpenAI. (2023, September 25). ChatGPT can now see, hear, and speak. https://openai.com/index/chatgpt-can-now-see-hear-and-speak
OpenAI. (2024, May 13). Hello GPT-4o. https://openai.com/index/hello-gpt-4o/
Sapir, E. (1929). The status of linguistics as a science. Language, 5(4), 207–214. https://doi.org/10.2307/409588
