- The need for English language learning (and other languages) is still there, potentially growing. AI (such as industry leader Cathoven AI) hasn’t taken that away.
- But when countries reduce their intake of international students, those students worry about completing a four-year degree, so demand (for ESL/EFL teaching) wanes.
- The global economy also challenges the language education industry; English language study abroad is less affordable for most middle-class families, world-wide.
- The Pandemic has similarly affected students’ English studies. It makes sense to “stay home to stay healthy” when learning a language. and as Tocker says at the end of 2025, “Duolingo is having a great year” teaching students online.
- AI can improve language learning by lowering the stakes when giving students in-time (individual) feedback, AI gives confidence to students to try speaking, when they’re not in front of many peers (“a safe comfort zone”), or by placing them in different, simulated settings. And hiring an AI teacher is cheaper than working with a live tutor, over the same number of hours.
- But some things are lost when language teaching goes online: AI can give “too much feedback,” consistently, which can make students feel there’s no end to the need for improvement. By contrast, a human class offers a (provisional) end, so learning can coalesce in students’ brains.
- AI also can’t provide the context by which students’ brains process and learn new things. Only a classroom can provide an “experience.”
- Tocker says we must ensure our education systems develop 21st Century skills, including “soft skills” (e.g. workplace readiness, but the “workplaces” of the future are “fuzzy” now). GenZers will need to learn how to network and handle job interviews. When they’ve spent all their time using AI, they may lack such “soft skills.” Who will teach them those?
- Community and context are very important (e.g. both Long and Tocker met at a live [in-person] conference and their online conversation spun out of that in-person meeting).
- A useful analogy is MS Excel: when Excel was invented, it didn’t end the teaching/learning of mathematics. But Excel provided a tool that freed specialists from using pencils and paper.
- Excel and AI are technologies that humans now can use.
- But AI is (of course) more complex than Excel–it will take much more time to figure out how to incorporate AI into education and all vocational fields (e.g. accounting).
- AI testing won’t replace standardized language exams, like IELTS. But Tocker says it will “shrink the number of players” in the space of English language testing.
- There are many limits to standardized language exams. Students often worry more about learning exam-taking skills than they do about learning how to communicate accurately. ESL should never take as its focus only standardized exams. (Teaching students strategy to master a particular kind of test is not ultimately edifying.)
- Human teachers can help students to improve intonation, learn more collocations and impart students with skills needed in life.
- One way to empower language education (including great teachers) is to invest some of the profit from (language testing) companies to sponsor students from “have-not” countries. That investment would help students to gain access to overseas colleges and companies, where they can learn new languages.
- Over time, as Tocker concluded, “patient” strategies for teaching move education and the workplace ahead, better and faster, than “top-down,” hierarchical approaches. But enlightened education requires patience and won’t develop and evolve as rapidly as AI does.