Taiwan is moving fast to bring AI into education, especially English learning. The Ministry of Education (MoE) is setting up public platforms, while the commercial sector is promoting private AI tools. But both efforts are leaving a larger task underdeveloped: building the human infrastructure around the technology. Taiwan needs teachers who can engineer learning and students who can use AI without offloading memory, judgment, or the ability to speak and think with other people.
At a Taipei event on June 15, cram-school associations, EZTalking, Kang Hsuan Educational Publishing Group (康軒文教集團) and smart-classroom vendors signed an AI English Learning Platform MOU. Their goal was a smart English learning ecosystem built on AI tools, teaching content, smart displays and classroom recording systems.
The MoE’s Digital Learning Enhancement Plan put NT$20 billion into digital content, devices, school networks, and education data systems from 2022 to 2025. Its AI Talent Ark Project (人才方舟計畫) calls for AI learning environments, cross-disciplinary AI teaching talent and stronger data-based decisions.
Photo: Lai Hsiao-tung, Taipei Times
For English teaching, the platform Cool English had over 2.7 million registered users by the end of March, while about 65,000 teachers were using it for lesson preparation and teaching. Its tools include speaking bots, scenario-based conversation practice, pronunciation assessment and writing support.
Taiwan now has tools, funding and private-sector speed. The weak point is learning design that can help students develop what the MoE calls “abilities AI cannot replace”: insight into the real world, the ability to use AI without being limited by it, collaboration, empathy, creativity and human-centered decision-making.
‘LEARNING ENGINEERS’
Photo courtesy of Taiwan Cement
For most of history, a teacher’s authority came from knowledge. Teachers knew more than their students and could explain what students could not yet understand. That advantage is fading. AI can explain grammar, translate sentences, create examples, correct writing, generate summaries and give pronunciation feedback — better than most teachers can.
Yet a speaking bot or smart screen cannot build classroom culture, read hesitation, create trust or help a shy student speak to another person.
Teachers therefore need to become learning engineers who design the conditions for attention, effort, feedback, memory, communication, collaboration, judgment and creativity. The classroom should give students experiences that turn short-term understanding into long-term memory.
Photo: CNA
Teachers trained in learning theory and AI tools can produce stories and songs that recycle textbook language in memorable ways. AI can turn a vocabulary list into a short mystery, dialogue or song. The teacher adjusts the level, checks the language and builds the follow-up task. Students then listen, predict, perform, discuss and create their own version.
COMBATING COGNITIVE OFFLOADING
Tina Austin, a biomedical researcher and AI educator who has taught at UCLA, has updated the traditional educational hierarchy that moves from remembering facts to creating new work. Her key point is that when AI can produce a polished essay, dialogue or presentation in seconds, the finished product tells us less about what a student has learned. Teachers need tasks that make students’ learning visible.
A weak AI English task asks students to generate a travel dialogue. A stronger task asks them to write their own version first, compare it with an AI-generated one, revise their work, explain the changes and perform it with a partner. Students use AI while keeping control of the judgment.
This protects against cognitive offloading. A MIT study last year led by researchers at the Media Lab found that participants who used ChatGPT for essay writing showed weaker neural connectivity, poorer recall and less ownership of their work than participants who wrote without AI. But when AI was used after students wrote their own versions, their recall was good and, importantly, the quality was significantly better. So, when students completely offload memory, composition and judgment, they learn less.
UNIQUELY HUMAN ABILITIES
English classes are well placed to develop the uniquely human abilities named by the MoE. Since AI can draft, correct, expand, and polish writing in seconds, English education should still test the fundamentals, but with some AI-mediated reading and writing while giving more time to AI-free speaking and listening.
Students speaking with classmates must listen, notice tone, read faces, respond to surprise, repair misunderstandings and adjust their words. They need empathy, problem-solving, judgment and the ability to think on their feet. In this way, speaking can turn critical thinking into social action.
While AI can supply vocabulary, models, feedback and extra practice before or after a task, students can solve a travel problem in pairs, negotiate a plan, role-play a complaint, defend an opinion or finish a story together.
Since students cannot pause every live exchange to ask a chatbot what to think or say; they must draw on memory, read the other person, make a choice and respond. These small acts of effort help resist cognitive offloading.
The public sector is installing the hardware and the private sector is loading the apps. Taiwan now needs the human infrastructure: teachers who can engineer the learning system and students who can use it to become more linguistically fluent and cognitively resilient.
June 29 to July 5 The daily morning flag raising ceremony at Laosong Elementary School (老松國小) was once a sight to behold, with thousands of students lined up on the main field. Boasting a record 11,110 students in 1966, the school wasn’t just the most populous primary institution in Taiwan — most sources claim it was the world’s most attended. At a time when students had to pass entrance exams to attend junior high, Laosong became a magnet school due to its high advancement rate. According to news reports, 12 buses transported students to the campus in Taipei’s Wanhua District
“You have several different internship experiences across different industries,” the interviewer told Hazel Lin (林敏淳). “But they are still internships, not jobs. Why should we hire you? What are you actually capable of?” It was a blunt question. It is also the question increasingly faced by many young Taiwanese workers trying to enter the AI economy, which is a key factor in transforming Taiwan into an indispensable global “AI island.” Lin studied international business, worked in marketing, supply-chain coordination and sustainability events, received a Master’s degree in customer relationship management in Germany and was completing an applied AI training program in Taiwan.
“Restarting” cross-strait dialogue, contends Chinese Nationalist Party (KMT) Chair Cheng Li-wun (鄭麗文), is the only way to avoid “Taiwan becoming the next Ukraine.” Presenting Taiwan negatively as the “next Ukraine” has become a staple of pro-KMT and pro-China discourse. The alignment of “Taiwan as Ukraine” with People’s Republic of China (PRC) trope needs unpacking. It is part of a package of narratives that serve a variety of PRC propaganda needs. Restarting talks? We don’t have government to government talks because the PRC rejects them. We can have talks any time the PRC wants, it just doesn’t want to. Indeed, it is
For a nation facing a potentially existential military threat from an aggressive, hostile neighbor, Taiwanese have a curious relationship both politically and culturally with the armed forces that stand between them and invasion. That needs to change, but historical baggage is hard to shed. In 1993, a Democratic Progressive Party (DPP) lawmaker made some disparaging remarks about military veterans. Outraged, a Chinese Nationalist Party (KMT) lawmaker flipped over a table and came at him in a rage, fists swinging. The DPP lawmaker was hospitalized as a result. Video of the incident is available on YouTube. Today’s violence in the legislature
Leave a Reply