At a neighborhood clinic down the street, a doctor completes a diabetes risk assessment in 25 seconds — a task that used to take 20 minutes. At a FamilyMart, the store manager finalizes today's fresh food order in three minutes instead of fifteen. The customer service reply you got on LINE? That wasn't a person. Taiwanese people interact with AI dozens of times a day, and almost nobody realizes it.
A Doctor's 25 Seconds
Dr. Chen is a family medicine physician at a community clinic in New Taipei City. He sees forty to fifty patients a day. Starting March 2026, his workflow gained a new step: the moment he opens a patient's chart, an AI diabetes risk assessment system from the 健保署 (National Health Insurance Administration, or NHIA) automatically runs a complication risk score. A color block appears on the right side of his screen — green for low risk, yellow for medium, red for high. Twenty-five seconds.
Three months ago, the same assessment meant manually reviewing medical history, cross-referencing lab results, and consulting clinical guidelines. At least 20 minutes. Before that, most community clinics simply didn't do it — there was no time.
The system is called "AI-on-DM," the product of a five-year collaboration between the NHIA and Google Health, now deployed across roughly 20,000 primary care clinics islandwide. What once required 40 specialists working three weeks to screen 20,000 people now takes 1 hour and 24 minutes — a 14,400x increase in capacity. That same month, the NHIA rolled out a Gemini-powered health education assistant inside the "健康存摺" (Health Passbook) app, used by over ten million people. Taiwan may be the first country in the world to integrate a large language model into a national public health app.
Dr. Chen says the biggest change isn't speed — it's capability. A community clinic could never afford a specialist team for risk assessments. AI gave every clinic that ability overnight.
AI Inside the Convenience Store
Six-thirty in the morning. The manager of a 全家 (FamilyMart) opens a tablet. Today's recommended fresh food order is already on screen. Weather, foot traffic, yesterday's sales, nearby events — the system has factored it all in. She tweaks two items, hits confirm. Three minutes. Three years ago, this took fifteen minutes of experience and gut feeling.
FamilyMart introduced its AI fresh food ordering system in 2022. Ordering time dropped 75%. Fresh food waste fell by nearly 10%. Combined with their "友善食光" (Friendly Time) program — a 30% discount on items approaching expiration — the chain now prevents roughly 370 metric tons of food waste per month. 統一超商 (7-Eleven Taiwan) deployed its own sales-prediction ordering system across more than 7,000 stores, shifting inventory management from "intuition" to "data."
Taiwan has the second-highest convenience store density in the world — one store for every 1,500 people. When AI penetrates this channel, it doesn't just reach early tech adopters. It reaches everyone who walks in to buy an onigiri.
LINE: Taiwan's Invisible AI Gateway
AI adoption in Taiwan didn't trickle down from Silicon Valley — it grew out of LINE group chats.
LINE has 22 million monthly active users in Taiwan, a penetration rate above 90% — higher than Japan or Thailand. Heavy users open the app an average of 22 times per day. This isn't just a messaging app. For most Taiwanese, LINE is the primary interface through which they access digital services. Think of it less like WhatsApp and more like a combination of WhatsApp, a mini-app store, a payment platform, and a government communication channel, all in one.
By late 2024, LINE Taiwan had over 3.09 million official accounts. One-on-one message volume grew 45% year-over-year; notification messages surged 4.5x. The driving force behind this growth is AI: LINE's official accounts now come with built-in AI chatbot functionality. Businesses simply upload product data or a URL, and the system auto-generates FAQs and real-time customer replies. What used to require an engineer to set up a chatbot integration now takes a few clicks in a dashboard.
In October 2025, LINE Taiwan declared the "AI Agent Era" at its annual CONVERGE conference, launching MINI HOME as a new user hub that aggregates multiple brands' LINE MINI Apps inside the LINE Wallet page. In the first half of 2026, LINE is rolling out LINE Touch — tap your phone on an NFC tag and instantly launch a service page for coupons, ordering, or event registration, all in a single touch.
Translation bots are another everyday tool in the LINE ecosystem. Third-party bots like Ligo support over 100 languages and are widely used in tourism and international trade group chats. LINE's CLOVA voice recognition technology has been localized for Mandarin Chinese in Taiwan, capable of transcribing speech in noisy environments for meeting minutes and customer service voice analytics.
Taiwanese people interact with AI dozens of times a day. It's just that the AI hides behind LINE's interface, looking like an "auto-reply."
The Other Side of Medical AI: From Colonoscopies to Bone Marrow Slides
The NHIA's diabetes model is just the tip of the iceberg. 雲象科技 (aetherAI), in collaboration with National Taiwan University Hospital and Cathay General Hospital, developed a real-time AI polyp detection system for colonoscopies. It has received medical device approval from the 衛福部食藥署 (TFDA — Taiwan's equivalent of the FDA), with 95.8% accuracy and latency under 0.2 seconds. The training dataset came from over 3,000 patients and more than 400,000 images. Their AI bone marrow smear classification and counting system holds both TFDA and EU CE certification — a world first.
Taiwan's 全民健保 (National Health Insurance, or NHI) system — which covers 99.9% of the population — has accumulated over 20 years of comprehensive medical records. This data goldmine is now being re-mined by AI. Models built on complication severity indices certified by three major medical societies (including the Diabetes Association of the Republic of China) analyze de-identified NHI data to risk-stratify more than 2 million Type 2 diabetes patients. Very few countries on Earth have a single database of this scale and completeness available for training medical AI.
Drones and Sensors in the Fields
In a fig orchard in Douliu, Yunlin County, a farmer applied for a smart agriculture subsidy and installed an intelligent skylight system and smart irrigation. The result: crop loss dropped below 5%.
Stories like this are repeating across Taiwan. The Ministry of Agriculture's "Grain Industry Upgrade Program" has established 111 rice production clusters nationwide, deploying sensors and AI image recognition to monitor soil and weather data in real time and determine optimal harvest timing. Farmers can now "patrol their fields" by checking a phone app — no more walking under the blazing sun all afternoon.
Drones are another rapidly spreading entry point. Field tests show that drone spraying reduces application time by 95% and pesticide usage by 60% compared to manual labor. A farmer in Zhongpu, Chiayi used a drone to spray his loofah garden — two plots done in under five minutes. In rice paddy comparison trials, traditional manual spraying achieved a 63% pest elimination rate; drones hit 93%.
Local governments in Chiayi County, Kaohsiung City, and elsewhere offer smart agriculture subsidies covering sensors, environmental control systems, and drones — up to 50% of equipment cost, capped at NT$500,000 (roughly US$15,500).
But technology diffusion brings new problems. Pesticide drift from drone spraying is a recurring dispute. In Chiayi, a rice farmer hired a drone operator to apply herbicide; the chemicals drifted into a neighboring tomato farm, killing over half the plants. A court ordered NT$450,000 in damages. There aren't enough licensed drone operators with dual certifications. Regulations say no spraying above 3 meters per second wind speed and a 20-meter buffer from neighboring fields, but enforcement on the ground remains patchy.
The technology is ready. Institutions and talent are still catching up.
AI Tutors in the Classroom
The Ministry of Education's 因材網 (Adaptive Learning Platform) includes a generative AI learning companion called "e度" (eDegree), which guides students through concepts via dialogue, practice problems, and positive feedback. As of March 2025, more than 750,000 teachers and students have used it, with over 30,000 logging in daily.
Behind that number is a structural demand created by the 108 課綱 (108 Curriculum Guidelines) — Taiwan's national curriculum reform launched in 2019. The new curriculum emphasizes learning portfolios, self-directed learning, and inquiry-based learning, all of which require individualized guidance. But Taiwan's student-to-teacher ratio doesn't allow every student to have a personal tutor. AI fills that gap.
Generative AI is also reshaping lesson preparation. Teachers are no longer confined to textbook frameworks — they can rapidly produce diverse materials, generate practice questions, and plan customized learning paths. In the cross-disciplinary literacy instruction that the 108 Curriculum emphasizes, AI serves as a research assistant, quickly synthesizing knowledge across subjects and drafting lesson plans.
Private organizations like the 均一教育平台 (Junyi Academy — Taiwan's equivalent of Khan Academy) are also developing AI-assisted adaptive learning tools for primary and middle school students. At the higher education level, the Ministry of Education runs the "AI Technology and Application Talent Cultivation Program" (AITCP), building cross-university AI course maps.
The real challenge, though, isn't technology. When 46% of Taiwanese consumers have already used generative AI tools — and 69% of 18-to-25-year-olds have (per the Institute for Information Industry's MIC survey, Q4 2024) — schools need to teach more than "how to use AI." They need to teach "how to tell whether what AI produced is actually correct."
How Taiwanese People Feel About AI
According to the MIC survey from the Institute for Information Industry (資策會), over 90% of Taiwanese have heard of AI. The share who say they understand generative AI rose from 24% in 2023 to 40% in 2024. Nearly 70% believe generative AI has real potential.
But "heard of" and "trust" are different things. The two biggest anxieties among Taiwanese are over-reliance on technology and privacy. Government regulation, data transparency, and sufficient background knowledge are the three preconditions people cite before they're willing to use an AI product.
The enterprise numbers are more telling. KPMG's 2025 Taiwan Industry AI Survey found that the average corporate AI adoption score was just 32 out of 100 — stuck below 40 for three consecutive years. Nearly half of companies have not announced an AI strategy. The retail and trade services sector scored a dismal 13.7. Over 40% of companies report having no AI data governance measures in place.
In other words: Taiwanese consumers are already immersed in AI through daily touchpoints — LINE customer service, convenience store fresh food, the NHI app — but most enterprises remain at the "tool-assisted" level, lacking a coherent AI adoption roadmap. AI permeated from the consumer side up, not from enterprise strategy down.
In December 2025, the 立法院 (Legislative Yuan — Taiwan's parliament) passed the 人工智慧基本法 (AI Fundamental Act) in its third reading. The 20-article law establishes seven governance principles: sustainable development, human autonomy, privacy protection, cybersecurity, transparency, fairness, and accountability. The National Science and Technology Council (NSTC) serves as the central authority; the Ministry of Digital Affairs handles risk classification. It is Taiwan's first dedicated AI legislation and one of the few in Asia enacted at the "fundamental law" level.
Taiwanese attitudes toward AI are more open than Europeans' — but also more anxious. The openness comes from the fact that LINE, NHI, and convenience stores have already made AI routine. The anxiety comes from the fact that nobody has explained how those "automatically processed" decisions are actually being made.
An AI Island Grown From LINE Group Chats
Taiwan's AI story has a distinctive structure: it spread along existing everyday infrastructure, growing from the bottom up. LINE is the communications infrastructure. Convenience stores are the retail infrastructure. NHI is the healthcare infrastructure. AI grafted itself onto systems that Taiwanese people already trusted — and that's why it penetrated so quickly and so deeply.
National investment is ramping up. The Executive Yuan's "AI New Ten Major Construction Projects" plan commits at least NT$190 billion (roughly US$5.9 billion) over four years, covering semiconductors and computing infrastructure as well as AI applications in healthcare, agriculture, education, and other public-facing domains. TAIDE — a Traditional Chinese large language model developed by the NSTC — is already deployed in seven areas, including 神農 TAIDE (Shennong TAIDE, an agricultural knowledge retrieval system) and Taiwanese Hokkien language instruction for primary and middle school students.
But numbers only tell half the story. What will truly determine Taiwan's AI trajectory are the people on the ground: the FamilyMart manager pressing confirm at 6:30 a.m., the family medicine doctor reading an AI risk report in a community clinic, the Yunlin farmer patrolling rice paddies from a smartphone.
None of them would say they're "using AI." They'd say: "The system recommended it." "The app said so." "The machine calculated it."
AI in Taiwan is no longer a tech news headline. It's part of everyday language — it just doesn't have that name yet.
References
- 全家「三減政策」AI 系統每月助攻減少 370 公噸剩食 — foodnext, 2022
- 健保署聯手 Google：AI-on-DM 糖尿病精準衛教如何改寫 320 萬患者的照護模式 — tenten, 2026
- How Google and Taiwan Are Building an AI Blueprint for Public Health — Google Blog, 2026
- 資策會 MIC 生成式 AI 調查：46% 消費者曾使用生成式 AI 工具 — Institute for Information Industry (MIC), 2025
- LINE BIZ CONVERGE 2025 全台官方帳號超過 309 萬 — LINE Taiwan, 2024
- LINE Taiwan's Vision for an AI Agent Era — LY Corporation, 2026
- 雲象攜臺大醫、國泰醫取國內首款大腸鏡 AI 瘜肉偵測智慧醫材證 — Global Bio & Investment
- 台灣 AI 治理新紀元：《人工智慧基本法》通過後的產業實戰解讀 — Taiwan AI Academy, 2025
- KPMG 台灣產業 AI 應用趨勢與展望報告 — KPMG, 2025
- 無人機噴農藥效果如何？農友驚嘆 5 分鐘噴完 2 分絲瓜園 — 上下游 newsmarket
- 教育部因材網 AI 學習夥伴超過 75 萬師生使用 — Youth Daily News, 2025