“In the future, noone will have to learn languages,” my grandpa said and adjusted his ski goggles. “With the help of technology you’ll one day be able to speak and understand every tongue on earth.”
We were inching our way up over a steep ridge deep in the French Alps, the chair lift swaying and creaking in the icy headwind.
“Perhaps for translating single words,” I said, thinking of my recent failed attempts to use Yahoo’s Babelfish for my French homework. “But a machine will never be able to fully understand the intricate meaning of complex sentences.”
“That’s the fun part,” he said as we approached the station. “It doesn’t have to. It just has to be good enough to fool us into thinking it can.”
I lifted up the bar and we jumped off the lift, skidding over the compressed snow. “But it will never be real!” I protested and stepped into my snowboard, the binding making a satisfying click. Having just recently picked language studies as my major I wasn’t about to give in that easily. “After all, language is what makes us human. A machine will never write poetry!”
My grandpa smiled, grabbed his ski sticks and pushed himself towards the descending slope. “Mark my words,” he said, and shot down a piste marked red.
Rise Of The Machines
It’s been 23 years since we had this conversation. Nowadays everybody uses Google Translate or DeepL, and whipping out one’s phone while abroad to translate street signs, menus and everything else has long superseded the well-thumbed pocket dictionary.
More recently, since its release in November 2022 a new language based AI language model by the name of ChatGPT has taken the world by storm. Almost overnight it has turned into the bane of teachers and lecturers worldwide, as students increasingly rely on AI-generated answers for their homework and (unlike my futile Babelfish experiments) often get quite passable results.
Needless to say, the machines still don’t understand language in any conventional sense of the word. But they have become darn good at faking it. Good enough to pass prestigious graduate-level exams, such as the final MBA exam at poetry.)
Simulating Simulated Knowledge
In a recent article the heads behind writing software iA Writer mused on how AI will not just change the educational landscape, but may also eliminate a lot of “bullshit” tasks in business such as creating memos, writing standardized emails, cover letters, meeting summaries, etc., in short anything purely functional:
AI has made perfectly clear that, to a large degree, we all simulate knowledge and meaning. It is so good at simulating school and business language because a lot of our own understanding in both spheres is largely simulated.
That’s actually a remarkable point. And it makes you wonder: have machines become so good at imitating us only because we ourselves have become increasingly imitable, made redundant by our own standardized ideals?
The Future Of Language Learning
So what does all of this mean for language learning? If AI is getting better and better why even bother with flipping flashcards and brooding over declension tables, right?
Well, it depends …
Personally, I neither share the blind faith nor the pessimism around this technology. In my mind it’s just another tool, and like all tools it is only as “smart” as its user. But I’m fairly certain that within the next few years it will greatly change how we approach many text- or language-based tasks.
Teaching languages over the last 15 years, first with groups, then one-on-one and currently with my German story books, I’ve come across two general types of language learners which often have very different goals and perspectives. I’m going to call them the Utilitarian and the Connoisseur for the sake of this argument.
A) 🔨The Utilitarian: needs to acquire proficiency for a specific end, such as an exam, visa, work, etc. or just get certain tasks done, preferably as fast and efficiently as possible. Once the goal is reached, regular study efforts often fizzle out, until the next target goal is acquired.
B) 🍷 The Connoisseur: enjoys the process of studying a language, gleaning cultural insights, staying sharp, sometimes as part of a social activity like a weekly book club, etc. Often an open-ended process for the sake of lifelong learning and self-actualization.
These groups are of course not always mutually exclusive. For example you may need to learn the language for a citizenship test, but still greatly enjoy the process, or vice versa: maybe you’re studying just for fun, but still utilize the language while traveling or with friends and family.
Having said that, I do believe however that the former group will be able to benefit significantly more from tech-assisted language applications than the latter.
If you just “need to get by” in German or any language for that matter, using technology can be a great time-saver. Similar to how nowadays websites are automatically translated to your preferred language, allowing you to parse information faster, you can use these tools to help you write notes to your landlord, employer, doctor’s office, etc. And sure, it will sometimes make weird mistakes, but at least it gives you something to work with. There’s still a bit of a stigma around using things like GoogleTranslate, and the resistance to AI is often immense, but if you know how to make these tools work for you responsibly (!), why not use them?
If however you are more in the Connoisseur camp, I don’t think that AI in and of itself will change how you approach language learning. Since for you, learning is not just a means to an end, you may actually want to take your time, let new vocabulary sink in, enjoy the sound of new words and smell the proverbial flowers along the way.
With all that said, I don’t want to imply that any of these types is better. Sometimes you just need to get things done. And that’s okay. And at other times, there’s more space to luxuriate in a language and really take in all the cultural influences.
Personally, I’ve always leaned more towards looking at language learning as an open-ended project. I don’t think you can ever “get done” learning a language, and if you closely pay attention you will discover deeper and deeper layers even in your own native tongue.
And perhaps as new technologies will slowly make more and more of our daily drudgery irrelevant, as more and more time will slowly free up, more people will explore this decelerated approach to language learning.
That’s certainly one way how I always looked at my German learning books. These are not really made for people who just quickly want to ace their Goethe A1 Test, although they may still help. Rather, I’ve always designed these books as a way to get people excited about the language itself, gain curious cultural insights and just have fun with mind-boggling idioms or zany characters.
Recently I made a little experiment where I tried to replace myself with ChatGPT, letting it write little German stories, come up with vocabulary and comprehension questions, and while it generally did an okay job, the outcomes were somewhat lackluster and lacking personality.
And perhaps singer-songwriter Nick Cave was right in his recent scathing review of AI-generated lyrics when he wrote:
“ChatGPT has no inner being, it has been nowhere, it has endured nothing, it has not had the audacity to reach beyond its limitations, and hence it doesn’t have the capacity for a shared transcendent experience, as it has no limitations from which to transcend.”
To which my grandpa probably would answer: “Dear Nick, AI doesn’t need a soul, it just needs to simulate it well enough.”
This article originally appeared on the LearnOutLive newsletter