Before Christmas, TechCrunch reported that OpenAI (the creators of ChatGPT) had raised $78M for an AI-powered language learning app called “Speak“. The app’s been around for eight years and has an undisclosed number of users learning English. Now, it aims to expand to other target languages, starting with French and Spanish. The app’s founders describe it as “part learning method and part tech platform”.

I haven’t tested this app out myself yet. But according to the article, what’s unique about this app is that it aims “to help users learn languages by talking out loud”:

Languages are typically taught by first exposing you to reading and writing, but native speakers always start learning languages by hearing and talking. Speak has built a platform to teach languages by focusing on how native speakers learn: Using AI, the startup generates audio conversations and listens to users’ responses to improve their grasp of a language.

Yet another second/additional/foreign language learning app evoking the first language acquisition process! My reaction to this was firmly 🙃.

I beg tech companies who delve into the language education space to please hire some applied linguists and/or language educators or researchers or ANYONE familiar with how languages are learned and taught.

It’s not untrue that “native speakers always start learning languages by hearing and talking (sic)”, it’s just that first language acquisition and second/additional/foreign language acquisition are very different processes.

In other words, your first language(s) (L1), the language(s) that you learn in childhood that you could be said to speak “natively” (or that could be considered your mother tongue(s) in many cases) are learned via one set of processes. There are different theories around the specifics, but in a nutshell, common to them all is certain conditions in a young child’s brain coupled with a necessity to be able to communicate. These spur the rapid acquisition of your first language.

Languages that you may learn later in life, at school or as an adult (i.e. second, additional or foreign languages) are learned differently for a large number of cognitive and sociocultural reasons. The brain is different in adulthood, and also the reasons, motivations and circumstances behind your learning of the language all influence how effectively we learn.

So although there are some commonalities between the processes, comparing first and second/additional/foreign language learning is a little like comparing apples to oranges. So why do tech companies keep promoting their second/additional/foreign language learning apps by making the compmarison to how people learned their first language?

Cases in point: Rosetta Stone says it “mirror[s] the way you first acquired language as a child” through “dynamic immersion”. Duolingo says it “allows learners to discover patterns on their own without needing to focus on language rules — the same way you learned your first language as a child.”

Other apps also evoke the same connection to first language acquisition processes and childhood. And now Speak.

I get that maybe they’re trying to contrast their apps and language teaching approaches to others which may be more text-based and/or rely on grammar-translation methods, or explicit explanations of grammar rules, etc.

But perhaps they’re also trying to justify the lack of explicit explanation of rules, usage or principles in their apps, by evoking the “immersion” or “rote” learning of “native speakers”. But really, is this just because they don’t want to hire teachers, curriculum experts, institutional designers or anyone else with pedagogical training?

It’s much less labour-intensive, and therefore cheaper, to just present learners with a bunch of random input and just let them “figure it out” under the guise of “implicit learning” or “immersion”. Sequencing and scaffolding input and carefully designing learning activities so that they’re difficult enough to push learners’ limits but not so hard learners get frustrated and demotivated (i.e. in the zone of proximal development or ZPD) takes expertise. So does writing good explanations of grammar, usage and other principles of the target language.

And expertise costs money.

It’s not even that I think that any of these apps are bad. I use lots of different tools in my own language learning as well as teaching, and the most successful learners usually have a personalized “cocktail” of learning tools which may include traditional language classes (group, or private; online or f2f), self study via books, apps or other tools, and some to type of immersion/applied situation such as travel or tours, conversation clubs, language exchanges online or f2f, etc. And these apps can definitely have a role in this.

So ideally any/all of these apps would stop using “immersion” as an excuse for cheaping out, and start hiring language education experts (I’m available, lol!) to complement the content in their apps with explanations and principles. Let users be able to toggle them on/off so that those who find explanations overwhelming can give them a pass, but those who prefer to understand what they’re learning can access that info.

That’s the beauty of second/additional/foreign language acquisition: children learning their L1 don’t need/want explanations of what they’re learning but some adult learners can benefit from this explicit understanding. I certainly do!

Post Script:

I have a LOT to say about some other aspects of this app, but that will have to wait for another post!

Zwick describes Speak as part learning method and part tech platform that works in a three-step process. 

First, you are thrown into listening and talking — an interesting approach, considering that Hsu and Zwick met and started working together after going through a cohort as Thiel Fellows, where you are, in theory, thrown into building an enterprise rather than going through years of learning first.

“We’re not going to explain every single grammatical rule to you,” Zwick said of the first step of its program. 

Second, you are then asked to apply that new term or phrase over and over — “basically drills where you just practice saying it out loud with various other pieces of languages so that it becomes automatic, with no translation.” 

Third, Speak then presents the phrase in “a real-world context using AI… That’s how you really anchor it,” Zwick said. 

Ironically, although the aim is to get its learners speaking a new language with humans, there are no humans in the loop. This is all crafted using speech recognition, natural language processing, generative AI and more to tailor the learning to the learner.