EdTech Today

OECD Digital Education Outlook 2026: Key Insights Will using generative AI help us study better? The OECD's March 2026 report, "OECD Digital Education Outlook 2026: Exploring Effective Uses of Generative AI in Education," answers this question with a resounding "not so simple." This flagship report, a synthesis of global empirical research, design experiments, and expert interviews by the OECD's Centre for Education Innovation Research (CERI), offers this key message: Generative AI has the potential to fundamentally improve education. However, depending on how it's used, it can also harm learning. We've summarized the key insights that run through the entire report. The "performance-learning mismatch" phenomenon, where assignment scores rise but skills decline. The most shocking study in the report is a randomized controlled trial (RCT) conducted in Turkey with 1,000 high school students (grades 9-11) . Students were divided into three groups and practiced math for six sessions (90 minutes each). Group 1: Self-study using only textbooks and notes Group 2: Using the general-purpose GPT-4 chatbot (GPT Base) Group 3: Using the GPT-4 chatbot (GPT Tutor) configured for educational purposes. During practice, the GPT Base group's accuracy rate was 48% higher than that of the self-study group , and the GPT Tutor group's was a whopping 127% higher . But what happened on the test without AI? The GPT Base group actually scored 17% lower than the self-study group . The GPT Tutor group also remained at a similar level to the self-study group. The OECD describes this as a "misalignment between task performance and genuine learning ." While AI provides answers for students, improving practice scores prevents the cognitive effort of students' own thinking, preventing actual knowledge acquisition. This phenomenon is also confirmed by neuroscience research. In an experiment where students from five American universities were asked to write essays, only 12% of those who used ChatGPT remembered what they had written an hour later, compared to 89% of those who wrote their own essays. Brain imaging analysis revealed that the AI ​​group's brain activity shifted from "generating" content to "supervising" the AI's output, and they showed significantly lower neural connectivity and engagement. Educationally designed AI is definitely different. So, does that mean AI is useless in education? The OECD answers unequivocally, "No." The key difference lies in the difference between using the general-purpose ChatGPT as is and using AI designed for educational purposes. Harvard Physics Class RCT: Students who learned online with an AI tutor implementing active learning principles achieved significantly higher learning outcomes, spent less time learning, and were more motivated and engaged than students who received the same teaching method in-person (effect size 0.63). Stanford Tutor CoPilot: This tool fine-tunes GPT-4 with feedback from top tutors and integrates it into an online tutoring platform. It was used by 900 tutors to teach 1,800 low-income students. Less-experienced tutors saw a 9 percentage point increase in their student passing rates , while tutors with previously low ratings saw a 7 percentage point increase . In contrast, there was no significant difference for previously high-performing tutors. A Chinese Problem-Based Learning (PBL) Experiment: Using GenAI in a problem-based learning approach in reading education resulted in improved reading performance and motivation compared to traditional, non-personalized methods. There's a principle consistently emphasized in the report: only AI tools designed or configured based on proven pedagogical principles, such as Socratic questioning, scaffolding, and active learning, will produce effective learning outcomes. Even general-purpose AI can be effective when used with "educational intent."

The use of AI in education is rapidly increasing. Tools like ChatGPT have ushered in an era where personalized learning support is available anytime, anywhere. However, one important question remains. "Does studying with AI really improve your skills?" To answer this question, OpenAI, in collaboration with the University of Tartu in Estonia and the Stanford University SCALE Initiative, developed the Learning Outcomes Measurement Suite. Today, we'll briefly explain why this tool is necessary and how it works. Limitations of Previous Research: Test Scores Alone Are Insufficient Until now, most research on whether AI helps learning has focused on test scores . If scores go up, it's effective; if not, it's not. But if you think about it for a moment, this is too simple. If you improve your score by cramming right before the test, is that really learning? What if your short-term memorization skills have improved, but your ability to think for yourself has actually decreased? Just because something works in one subject, does it mean it will work in others? Ultimately, it was difficult to answer deeper questions such as “Do skills last over time?” and “Do you develop the ability to study on your own?” using existing methods. OpenAI's First Experiment: Validating the Study Mode Effect OpenAI first conducted its own experiment , using AI study mode to help over 300 college students prepare for exams in neuroscience and microeconomics . Study Mode is an AI learning tool designed to go beyond simply providing answers, guiding you step-by-step like a teacher, checking your understanding, and providing practice opportunities. The results of the experiment were as follows. Microeconomics : Students who used study mode scored about 15% higher than those who studied using traditional methods (Google searches, YouTube, etc.) . Neuroscience : While the trend was positive, it was not statistically significant. Results varied across subjects, and the extent to which AI was utilized varied across students. This experiment convinced the OpenAI team that the true learning effectiveness of AI cannot be determined by simply looking at one or two test scores . A New Approach: What is a Learning Outcomes Measurement Tool? That's how the Learning Outcomes Measurement Suite was born. This tool focuses on three main areas: 1. How AI is teaching We assess whether the AI ​​model is behaving in a way that aligns with good teaching practices. We check whether it simply provides answers or encourages students to think for themselves.

Darren Mowry, who oversees Google's global startup support, has issued a warning to the AI ​​startup industry, stating that two business models are "on alert . " Limitations of the repackaging model The first is a model that repackages and sells existing AI . AI like ChatGPT or Gemini does the actual work, while startups simply add a pretty screen over it and sell it as if it were their own product. Apps like study assistants and writing assistants are prime examples. The second model is one that brings together multiple AIs in one place . These services advertise, "You can use GPT, Claude, and Gemini all at once." As AI companies expand their capabilities, the need to go through intermediaries is disappearing. Mowry says: "The industry is no longer accepting that AI models do all the work, and then you repackage it and sell it." A similar situation occurred in the late 2000s. Numerous startups resold Amazon Cloud (AWS), but when Amazon developed its own enterprise tools, most of them disappeared. The only ones that survived were those that developed their own expertise in security or operations consulting. A similar trend is felt in education. This isn't just a tech industry issue. Similar trends are also occurring in education. On February 19th, Google officially introduced Gemini-powered AI feedback to Classroom. When students submit writing assignments, the AI ​​creates a personalized draft of feedback, which teachers then review, refine, and deliver. In fact, this functionality was pioneered and successfully tested in the field by edtech tools like Brisk Teaching. Brisk is a powerful tool offering over 30 features, including AI feedback, batch grading, and the ability to review student writing progress. However, Google's integration of this functionality directly into Classroom, without requiring separate installation, recreates the familiar pattern of a large platform company entering a field pioneered by a smaller company. There are also administrative aspects to consider. For Korean teachers, working with HWP documents has always been a time-consuming task, and AI-powered services have made this process much easier for many teachers. At the same time, general-purpose AI tools like Claude and Gemini are becoming capable of directly reading and processing HWP files, making it crucial to consider what unique value each tool can add going forward. "AI creates problems for you," "AI creates lesson plans for you," "AI provides feedback for you"—the sheer number of educational AI tools has grown. As teachers gain more hands-on experience with AI, and platforms like Google begin to incorporate AI capabilities, the differentiation of each tool will become increasingly important. Ultimately, it's a question of depth Citing examples like Cursor (a coding specialist) and Harvey AI (a legal specialist), Mowry argues that tools that delve deeply into specific fields and build their own expertise and data are distinct. The key isn't "using AI," but what's built on top of it. Education ultimately faces the same questions. The more deeply a tool understands and incorporates these elements—curriculum, student learning flow, and the context of the school—the more likely it is to endure. In the AI ​​era, the value of an educational tool will likely be determined not by its use of AI, but by its deep understanding of education . Source: Rebecca Bellan, "Google VP warns that two types of AI startups may not survive ," TechCrunch, February 21, 2026

If you look at classrooms these days, there is one thing in common. The point is that students are already using AI. We use AI to organize assignments, write essays, and even prepare for discussions. Can this trend be stopped? It's virtually impossible. So, you have two choices: Prohibited Teach Estonia chose the second. And that choice led to a national project called **AI Leap (TI-Hüpe)**. Don't just introduce AI, teach how to use it. AI Leap is not a declaration to “put AI in schools.” AI that is already in place It is a movement to design how to deal with it . The core question is simple: How do we prevent AI from thinking for students? How can we ensure that AI doesn't weaken our ability to think? What role should teachers play? This project focuses on learning outcomes rather than technology. We use AI, but we design the structure so that students' thinking becomes deeper. Teacher-centered structure Here's why AI Leap is exciting.

Key Summary Assessment was a hot topic at the London Bett exhibition in January 2026. AI scoring, AI feedback, and AI diagnostic solutions were abundant at every booth. However, the conclusion from the field was not that "AI replaces assessment," but rather that "AI broadens teachers' eyes and hands . " Professor Hannah Fry likened AI to "fast food" and warned that if AI takes away students' ability to grapple with the process themselves, education itself will disappear. Major companies like Magma Math, RM Compare, Turnitin, Redmenta, and Olex.AI all designed their AI to "assist teachers in making decisions," rather than "determining grades." While AI excels at grading math, classifying misconceptions, and speeding up feedback, it has structural limitations in assessing creativity, context, ethical judgment, and soft skills. The real question is not "Can AI replace assessments?" but **what can AI take off teachers' shoulders?"**. 1. At the exhibition hall where 30,000 people gathered, the hottest word was 'evaluation.' From January 21st to 23rd, 2026, Bett 2026, the world's largest edtech fair, took place at ExCeL in East London. Over 30,000 educators, school leaders, and technology executives from around the world gathered, and hundreds of booths showcased their latest technologies. This year's official theme was "Learning Without Limits." But as I walked around the exhibition hall, a different word dominated the space: 'Assessment .' Solutions using AI to grade, provide feedback, and assess student performance were visible in every hall. It's rare for "assessment," rather than hardware or robots, to be so prominent this year. Even the Bett Awards created a new award category dedicated to assessment. Sector Premier Note AI for Teaching and Assessment Aristotal AI-based integrated teaching and assessment solution Assessment Planning & Progress Monitoring Olex.AI Handwriting recognition + automatic generation of personalized feedback Company of the Year (< £3m)

It's the same as organizing the refrigerator. Have you organized your refrigerator? Take everything out first Throw away expired items I'll leave you with just one similar one Put it back in EdTech is the same. The problem is that I clean out my refrigerator at least once a year, but I never clean out my edutech. Things to know before organizing Why don't you organize it? Reason Inner feelings "I thought I might use it someday" Not used for 3 months "It's free, so there's no harm in having it." Login information is distributed "I don't have time to organize" There is time to manage unused tools. "I don't know who wrote it" If you ask, no one uses it What happens if you don't organize

Key takeaways from Leon Furze's "Teaching AI Ethics" (2026) It's been over three years since ChatGPT was introduced. In the meantime, AI has become deeply embedded in our daily lives. Microsoft 365 has Copilot, Google Workspace has Gemini, and new iPhones have AI built into Siri. Meta has deployed its Meta AI on every platform, whether users want it or not. But as technology becomes more widespread, ethical issues become more serious. Leon Furze began his AI ethics training materials as a blog series in 2023 and updated them into a book in 2026. He directly refutes the argument that "we should stop discussing ethics because AI is already ubiquitous." His core argument is that the ubiquity and inevitability of AI make ethics training even more crucial . Let's examine each of the nine ethical areas covered in this book. 1. Bias: Three Layers of Discrimination AI bias isn't simple. It involves three layers. Data bias is the most widely known problem. Training data scraped from the internet overrepresents certain groups—typically white, male, and English-speaking. This is the "probabilistic parrot" problem, warned about by Emily Bender and Timnit Gebru. AI simply mimics patterns it sees in the data. Model bias is more subtle. Research has shown that even GPT-4o, which has undergone safety training, exhibits cultural bias when given indirect prompts. The model doesn't understand "fairness." It simply learns and sometimes amplifies patterns. Human bias creeps in during data generation and labeling. Even Fei-Fei Li, the founder of ImageNet, admitted to being shocked by the racist and sexist labels in his dataset. If you want to see for yourself how serious this problem is, try creating a "CEO photo" on Midjourney. You'll see a bunch of white men in suits. Create a "nurse photo" and you'll see images of sexually objectified women. ChatGPT and Microsoft Copilot offer more diverse results by using guardrails like system prompts. However, these are merely Band-Aids . The underlying data bias remains. 2. Environment: AI is a mining industry. To borrow Kate Crawford's phrase, AI is an **extractive technology**. Data centers consume 3-4% of the total energy consumed in the United States. The water used for cooling is literally depleting water resources in some areas. Mining of minerals like lithium and rare earth elements, as well as the short life cycles of electronic waste, are also serious issues. It's even more shocking when you look at the numbers. GPT-3 training: 1,287 MWh of electricity , approximately 552 tons of CO₂ emissions This is equivalent to the annual electricity use of 120 average American households. Energy consumption at inference time (when we actually use AI) is also significant. Image generation consumes dozens of times more energy than traditional AI tasks like text classification. As educators, we must ask ourselves: "Do we really need ChatGPT for this task, or would a more efficient tool suffice?" 3. Truth: Hallucinations are a feature, not a bug.

"Because it looks good" is not the reason I saw it at a training session. I saw a demo at a trade show. A fellow teacher recommended it. So we introduced it. 3 months later. "What was the password?" To break this pattern, we need to ask questions before we introduce ourselves . Question 1: "What would be inconvenient without this?" This is the first question you should ask. ❌ Reasons for incorrect introduction "I heard everyone uses it these days." "It has AI capabilities." "It's free." "The demo was great." ✅ Reasons for correct introduction "I can't see the student's solution, so I don't know where they're stuck" → Snorkl "It takes too long to write feedback" → AI feedback tool "I don't have time to create class materials" → Canva Test : If you can't say in one sentence, "I'm uncomfortable without this tool because _____," don't introduce it. Question 2: "Is the time you save more than the time you learn?" Learning new tools takes time. Do the math

"It's in those painful moments, that's where the real learning happens." - Professor Hannah Fry Last January, 37,000 people gathered at ExCeL in London for BETT 2026 , the world's largest education technology exhibition. There were 700 booths and educators from 130 countries. I was among them. There is something I realized after walking endlessly for three days. It wasn't, "We need to introduce one more new edtech," it was, "We need to reorganize three of the ones we're using now." AI is fast, but nutritious, like fast food. Mathematician Professor Hannah Fry took the stage for the opening . The words she spoke really resonated with me. Junk Food AI It's fast and convenient, like a McDonald's hamburger. The answer is available with the press of a button. But it lacks nutritional value. The answer appears on the screen before the student can even ask, "Why?" Professor Fry said: "The moment you start to rack your brain and groan, that's where the real learning happens." We try to make things "easy" for our students. But the moment we take away that "difficulty," the learning disappears as well. UK Education Secretary Bridget Phillipson also nailed it: "Can AI replace teachers? Absolutely not. But it can free teachers from chores and free them to focus on teaching." The role of technology has become clear: not to replace, but to liberate . EdTech is piling up like expired sauce bottles in the refrigerator. Here are the hottest topics at BETT 2026. A phenomenon dubbed "edtech fatigue" by the UK's Education Policy Institute (EPI) . What does it mean? Open the fridge door. Ten half-empty bottles of sauce. Five jams that make you wonder, "When did I buy this?". Three expired dressings. School computers are the same. EdTech that I signed up for after receiving training last year (I tried it once and forgot about it)

A free AI ethics education tool created by Harvard Graduate School of Education. A framework for students and teachers to jointly define the scope of AI use. Visualized as a spectrum from "completely okay" to "over the line" Voting apps allow you to gather opinions from large groups. Currently in prototype stage, available in English only. The most troubling question in schools these days is: "Teacher, can we use ChatGPT for this assignment?" Simply saying "no" is too ubiquitous, as AI has become a commonplace feature, but saying "use whatever you want" obscures the learning objectives. How should we navigate this ambiguous gray area? Graidients, developed by a research team at the Center for Digital Thriving at the Harvard Graduate School of Education, aims to address this very issue. The name itself is intriguing: a play on words combining "gradient" and "grade." The goal is to create a tangible "gradient" in the "gray areas" of AI usage. The usage is surprisingly simple. First, choose a specific task, such as "Write a five-page essay analyzing the themes of The Great Gatsby." Then, brainstorm all the ways AI can be used for this task: grammar check, idea brainstorming, outline writing, draft generation, complete essay writing, etc. List everything without judgment. Here's the key step: placing each AI application on a spectrum. The left end is "completely okay," and the right end is "over the line." Students each place their ideas on the spectrum, compare their results, and discuss why some people think it's okay, while others don't. Finally, we reach a consensus on the specific assignment. It's important to note that this isn't a school-wide rule, but rather applies only to the specific assignment. Different assignments may have different standards. For example, AI tolerances don't necessarily have to be the same for essay assignments and math problem solving. There's also the Graidients Voter app, which can be used in large groups or remote classes. It allows you to vote in real time and view results at a glance, making it ideal for workshops and lectures. The greatest strength of this tool is that it engages students as rule-makers, not just objects of rules. Rather than simply prohibiting them with a "don't use it," it encourages them to think for themselves about why some uses are acceptable and others are not. Another major advantage is that it's free. Of course, there are limitations. It's still in the prototype stage, so its features aren't perfect. It's only available in English, so translation is necessary for immediate use in Korean classrooms. Furthermore, effective discussions depend on the teacher leading them effectively, so simply having the tool doesn't automatically make things work. Still, if you're a teacher who constantly struggles with the question, "Can we use AI?", this approach is worth exploring. It's not a tool that provides answers, but a framework that initiates a conversation to find the right answer together. Graidients Graidients

The real competition in edtech begins the moment you leave school. In the past, the path to education was simple. Kindergarten → School → College → Employment. And the training is over. But now it's different. Learning never ends, but rather It operates more fully outside of school . EdTech is at the heart of this change. But the important thing is, The point is that not all edtech is moving in the same direction . This article follows the flow of learning, At each stage This is a calmly organized record of what role edtechs actually play . 1. Learning no longer begins at school. The significant changes in the field of early childhood education are clear. Now the key is not ‘educational content’ How to connect parents, children, and the environment . Services like Lovevery Rather than teaching knowledge Structure developmental experiences . More than what you learn, The central issue is in what order and context it grows . At this stage, edtech has already It is closer to childcare and development infrastructure than an educational tool .

43% of American middle and high school teachers use AI detection tools. However, the error rate of these tools ranges from 15% to 45%. The situation is even worse for non-English speaking students. Nearly half of them may be unfairly punished. Things that actually happened Marley Stevens , University of North Georgia student. I used Grammarly, the university-recommended spellchecker. I was flagged for AI cheating. I received a zero grade. I was placed on academic probation. I underwent disciplinary action for six months. Ultimately, I lost my scholarship. "I couldn't sleep or concentrate on anything. I just felt helpless." Ailsa Ostowitz , 17-year-old high school student from Washington, D.C. I was questioned three times in one school year. Three times in two subjects. All three times, I wrote my own essays. "I'm mentally exhausted. This is something I wrote with my own thoughts." NSW public high schools in Australia. An AI detection tool incorrectly identified students' assignments as AI-generated. Students were publicly reprimanded by their teachers during class and even detained after school. It was all a false accusation. In numbers A 2025 study analyzed over 10,000 essays. Here are the results: Detection tools Misjudgment rate ZeroGPT 28% GPTZero 22%

The Christmas phone call that changed my life Last Christmas, Jo Muirhead, a primary school teacher with 20 years of experience, was enjoying her school holidays when the phone rang. "Would you like to talk to me about a new position at Canva?" We met once, talked, and that night, I got a job offer. Jo asked her boss for a year's leave, and now she's officially leaving teaching to promote Canva across Australia and the Asia-Pacific region. "I truly have the best job in the world." Megan Townes, also from the same team, took a similar path. After teaching middle and high school for 10 years, she moved to Microsoft Education Consultant 11 years ago and is now the Asia-Pacific Regional Director for Canva Education. "I tell teachers who are thinking about leaving the profession, you never know where you'll be in 10 years." Their story was featured on the TeachTechPlay podcast and is inspiring educators around the world. The Power of Canva by the Numbers 📊 Canva is currently used by 260 million people worldwide . Amazingly, more than a third of them are teachers and students, and it's free to use. But many teachers don't know this! Many worry, saying, "When students log into Canva, it says it's a paid version." The solution is simple: ask your school administration to sign up for Canva for Schools. It also integrates with Google Classroom and Microsoft Teams. Educators alone have created 1.5 billion works . And over the past few years, AI functions have been used more than 22 billion times . These figures are from a few months ago, so they're likely much higher now. Jo says, "Canva is a great place for students to safely learn and experiment with AI. We're preparing kids for the unknown future of careers. As teachers, it's our responsibility to keep up with AI and teach students how to use it effectively." The age of imagination has arrived. When Megan shared a quote from Canva founder Mel Perkins, many educators resonated with her. "We have entered the age of imagination . The people who will shine in the future will not be those who know how to use AI, but those who know how to use imagination ." Canva makes designing easy, but AI is just a tool. The real magic comes from your creativity. Canva's AI features: Magic Write : Creating and Editing Text Magic Media : Create photos, drawings, 3D images, and videos Background Remover : Erase background (oldest AI feature) Canva Code : Create interactive content like games and quizzes (essential!)

In the age of AI, a new group of entrepreneurs is emerging: "Teacherpreneurs," founders who are teachers with classroom experience. Transition from teacher to entrepreneur Jon Laven (Snorkl) I worked as a high school math teacher in Minneapolis for ten years. I started Snorkl with my co-founder, Jeff Plourd, a former teacher. My motivation for starting Snorkl was that I wanted math classes to be centered around students' thought processes, not just about getting the right answer. I worked part-time for a year and a half, developing the product while maintaining my full-time job, and tested it with students in real classrooms to get feedback. Jamie Poskin (TeachFX) I taught math and English at a private school in Harlem, and even served as the athletic director. While I was working on my MBA and Master of Education, a principal suggested I get the idea. "I'd love to get into the classroom, but I don't have the time. Video is great, but I don't have time to watch all the videos. Couldn't I use Stanford's great computers to analyze the recorded videos?" This was the beginning of TeachFX. Maya Bialik (QuestionWell) This is a case of a researcher-turned-teacher. After writing the book "AI in Education" in 2019, I became a teacher because I wanted to gain hands-on classroom experience. During February break, my husband was working on an AI-powered question generation tool, and I stumbled upon Maya, which did the same thing I'd been doing by hand. "This is something I have to do every day, so I can't help but automate it," he said, and he created the tool, which became QuestionWell. Merissa Sadler-Holder (Teaching With Machines) I've been teaching in the classroom for 13 years. Seeing the AI workshop at the 2023 ASU+GSV Summit so popular that I couldn't even get into it, I realized, "AI can be another burden for teachers. We need to create meaningful professional development and create spaces that amplify teachers' voices." What Teachers and Entrepreneurs Have in Common Quick adaptability Jon Laven: "Great teachers constantly adapt. To new curricula, new students, new situations. If a class isn't going as planned, they look at the data and quickly change course. Startups are no exception." Learning from repetition and failure Jamie Poskin: "How do you teach students who hate math and lack fundamentals? You can't hit a home run on the first try. You have to formulate a hypothesis every day, try it, find small successes, and build on them. It's the same iterative process as a startup." Comfort with uncertainty Maya Bialik: "I've been doing improv comedy for 10 years. It's taught me to be comfortable with uncertainty, to listen, to think on my feet, and to see mistakes as opportunities. That's what helps me as a teacher and as an entrepreneur." People management skills Merissa Sadler-Holder: "When I was transitioning into corporate work, I was told, 'You're just a teacher.' It's funny. People don't realize how much we do in the classroom. Managing different personalities, assessing motivations, resolving conflicts. It's something we do every day." Networking skills Merissa: "Most teachers aren't on LinkedIn because they don't have a need to network. But if you give teachers a voice and let them get out there, amazing things can happen. Networking is something we do every day in the classroom." How Your Teaching Background Affects Your Business Trust with customers Jon Laven: "When we talk to math teachers, we talk about educational philosophies like Peter Liljedahl and Thinking Classrooms. They say, 'Oh, I can see you've been in the classroom.'" Jamie Poskin: "I felt like I was talking to someone who understood the problem. It was 100% positive." Authenticity Merissa Sadler-Holder: "When I talk to teachers, I say, 'I'm one of you.' You're the kids who disrupted class every day, and every single gray hair on my head is named after you. I'm on stage, but I'm not here to tell you what to do. You already have the information, you're the experts. Teachers respect those who stand in line." Ability to switch contexts Maya Bialik: "It depends on who you talk to. They might say three years of classroom experience, or they might say a Harvard Masters degree and a published book. Both complement each other in important ways." Advice for Aspiring Teacherpreneurs Start small Jon Laven: "I started with Jamboard and Screencastify. I tested it in my classroom with real students. I didn't even plan on building an AI tool at first. I just watched the videos and gave feedback. Most ideas can be started by mimicking them manually." Just get started Maya Bialik: "Just give it a try. There's a lot of overlap between what teachers need and what entrepreneurs need. You'll have fun even if you don't know anything. You'll be learning along the way." Ask for help Merissa Sadler-Holder: "People genuinely want to help others. Teachers are reluctant to show vulnerability, but if you reach out to them on LinkedIn and say, 'Can you spare 10 minutes? I'd love to hear your advice,' people want to help."

Joo-Hyeon Joo, CEO of __T100009_____, introduced Redmenta and Snorkl during a GEG Daejeon webinar. The hour-long session covered two different AI training platforms: text-based and voice-based. A guidebook link was shared at the start of the lecture. This resource was designed to address the practical challenges of using YouTube lecture videos alone. It included a Redmenta guide and a snorkel guide. Redmenta's new direction Redmenta, a Hungarian company, is an education platform equipped with 160 AI-powered tools. Just two years ago, creating worksheets with AI in one click was a trend. Now, anyone can do it. Google also offers generative AI capabilities for free through Gemini. So Redmenta changed its strategy. Starting this year, conversations with AI and content creation are completely free. There are no restrictions. Instead, credits are only awarded for the evaluation function. While anyone can create content, the strategy emphasizes the importance of accurate evaluation and feedback. A system that uses three AI models simultaneously Redmenta utilizes all three AI models: GPT, Gemini, and Claude. This represents a shift in edtech AI beyond 2024. While previously relying on a single AI platform, Redmenta now tests all three models within a lab within the platform. GPT excels at math problems, while Claude excels at writing. The optimal AI is automatically selected and used based on the situation. The AI-based approach is no longer crucial. We've entered an era where the entire system is leveraged. Practical Use: Creating Custom Worksheets The main representative typed "custom" on the screen. Redmenta has 160 AIs, but too many can be difficult to find. That's why we use an AI-to-AI approach. When you search, it recommends the right tool. I uploaded a PDF file of a high school English textbook. The text was extracted, and I asked, "Let's design a plan for how to use this with first-year high school students." The AI suggested a structure for each lesson. However, instead of creating it right away, we could enter conversation mode and design it together. When I asked again, "Instead of just creating a worksheet, let's design a lesson together," the AI began asking step-by-step questions. It asked about learning objectives, core competencies, and other topics, and worked with the teacher to develop the lesson. Similar to ChatGPT, it specialized for education, providing more tailored answers. Audio generation and AI chat features When creating an English question, I requested, "Let me take an English listening test, talk to the AI, and finally write an essay." The AI created all the questions in English and even generated the voice. The sentence, "My family is very important to me," was played in a natural voice. You can create various types of audio, such as conversations between men and women or between friends. It works in all languages, not just English. It's also downloadable, so you can use it for other purposes. You can even embed an AI chatbot within a worksheet. When a student types "Hello," the AI initiates a conversation in English. If a student asks in Korean, it responds in Korean as well. Teachers can customize the AI's behavior. If a student wanders off topic, it can prompt them to return or explain why. While some teachers may have used chat-based tools like Mizou, Redmenta differs in that it allows you to embed worksheets within the chat. You can even create a chat and give it to students. It's a conversational learning approach. Creating a Feedback Character Before assigning assignments to students, teachers can define the nature of AI feedback. By specifying specific details like "Always praise students for their work, kindly address areas where they need help, and end with a cheerful tone," the AI will provide feedback in that style. The test results yielded feedback along the lines of, "Great! I'm so impressed by how hard you worked on task 2." The AI embodies the teacher's character. Personalized learning and group management

Program Introduction Apple Distinguished Educators are a global community of educators who leverage Apple technology to create innovative learning experiences. These leaders deliver personalized learning through iPad and Mac, inspiring other educators. 📋 Application Eligibility ✅ Eligibility Public/private school teachers University professor Non-profit organization/organization educator ❌ Eligibility for application School superintendent University President Top-level leaders with purchasing decision authority 🎥 Application Video Production Guide Essential Content (3 Topics) 1. Innovating the learning environment Case studies of innovative learning environments utilizing Apple products Specific methods that contributed to achieving learning objectives 2. Evidence of Success Data or learning outcomes that demonstrate impact Examples of student academic achievement or innovation strategies 3. Expand your sphere of influence Activities Beyond the Classroom/Campus/Community

"It was hard to find the class materials because they were scattered all over the place, right?" 🎯 Key points 3 ways to embed Canva : Upload an image, a PDF file, or a live link. Choose the one that suits your situation. Wakelet Custom Templates : Canva templates with the perfect background, header, and banner size. The advantage of live linking : Any edits you make in Canva will automatically be reflected in Wakelet. Automatically create quizzes : Create quizzes and check student answers with Canva Magic Write. Using Canva Code : Create Games and Drag-and-Drop Activities Without Knowing How to Code What is the Wakelet + Canva integration? At the 2025 Wakelet AppSmash Festival, teacher Jackie from New York and coach Frankie from Dallas, Texas shared how they use Canva in their classrooms. All Canva features are available directly within Wakelet, allowing students to access all their learning materials in one place. 3 Ways to Add Canva to Wakelet Method 1: Save as an image and upload After completing your design in Canva, you download it as an image file and upload it to Wakelet. You'll need to re-download and upload it each time you make changes. What are the uses? It's suitable for finished posters or brochures that don't require further modification. Method 2: Upload as a PDF file Download your Canva design as a PDF and add it to Wakelet. It will appear as a file link, and clicking it will open the PDF. You can also import directly from your Canva project folder by clicking the "Design with Canva" button in the Wakelet editor. Tip : Using the Canva integration saves you time by allowing you to select from your Canva projects without having to download them each time. Method 3: Connect via Live Link (Recommended) Paste Canva's shared link into Wakelet and it will sync in real time. Any changes you make in Canva will automatically be reflected in Wakelet. Practical use case : This is useful when collaborating with multiple teachers or students. While a Wakelet collection can be managed by a single person, Canva designs can be edited by multiple people. This allows collaboration without complicating the Wakelet collection. Teacher Jackie's Perfect Size Template

AI is changing the landscape of education - with one-third of the world's population offline, cutting-edge AI models are only accessible to those with subscriptions, infrastructure, and language advantages. Inequality is deepening – this divide determines not only who has access to AI, but also whose knowledge, values, and language dominate the systems that influence education. Philosophical, ethical, and pedagogical dilemmas exist – we explore the disruptive impact of AI on education, gathering insights from thinkers, leaders, and changemakers around the world. We need a global forum for dialogue - 21 papers explore everything from dismantling evaluation systems to fostering an ethic of care, creating a shared space to discuss across differences and reimagine inclusive education in the AI era. Building on UNESCO recommendations - the UNESCO Recommendation on AI Ethics, Guidelines for Generative AI in Education and Research, and the AI Competency Framework for Teachers and Students - with human rights, justice, and inclusion at its core. Step 1: Recognize Your Current Situation AI is entering education. A child in Ghana practices algebra with an AI tutor on WhatsApp. A teenager in the UK confides his anxieties to an AI "companion" who sends comforting messages 24 hours a day. At a university in South Korea, an AI avatar of a prominent professor conducts an entire class in multiple languages. These scenes, unimaginable just a few years ago, are now rapidly spreading across classrooms and homes around the world. But a serious contradiction exists Generative AI is showing mixed results. While personalized AI assistants are available to some, many still don't have access to them. While it promises speed and efficiency, it risks weakening critical thinking and exacerbating gaps in accessibility, gender, and language. It generates data on an unprecedented scale, but also raises unanswered questions about privacy, ethics, sovereignty, and trust. Step 2: Ask a fundamental question We need to redefine humanity AI is more than just a tool. It's fundamentally reshaping how we learn, teach, and understand the world. Ask yourself: → Who is a ‘student’ in the AI era? → Who is the ‘teacher’?

"The day before the exam, the student asked the AI, 'Will you give me a quiz?'" One student used a math homework helper app. The answers came out well, and the homework was done quickly. It was convenient. But on test day, the student completely failed. "I just got the answers, but I didn't understand them." When Google product manager Dave Messer heard this student's story, he realized something fundamental needed to change: AI needed to become a teacher who guided real learning, not just a butler who simply dispensed answers. AI knows the answer, but can't teach it. AI models can pass the bar exam, solve math Olympiad problems, and write doctoral dissertations. But teaching them is a different matter. "Imagine a robot sitting in a room," Messer says. "If you ask it what the capital of the Philippines is, it can tell you. If you ask it for a recipe, it can tell you. It can even pass a test. But if you say, 'Teach me something,' it just gives you the answer. It doesn't act like a teacher." Existing AI systems try to do everything for students, acting like butlers, providing them with the answers they need. But teachers don't teach like that. They ask questions, encourage students to think, and guide them to find their own answers. This is precisely where the core problem with educational AI lies. An AI model specifically designed for education, created by a former teacher. Google has been tackling this problem head-on since 2022. Rather than simply creating a smart AI, they set out to create one that "understands education." The result is LearnLM, a dedicated learning model. LearnLM is based on five educational principles: Encourage active learning - don't just give information, make people think. Cognitive load management - don't give too much information at once, but divide it appropriately. Strengthening metacognition - making people think about how they are learning Stimulate curiosity - it makes you want to know more, not just memorize. Learner-tailored adaptability - tailored to each individual's level and needs. Messer is a former teacher. His own classroom experiences and struggles are reflected in this principle. "You can't understand cell structure just by reading about it. It's only when you look at a diagram, ask questions, and connect the dots that you finally get that 'Aha!' moment." Gemini 2.5 Pro is the model that most deeply reflects these pedagogical principles. In fact, in a study targeting teachers, it was preferred over any other AI model. What Guided Learning Changes

📌 Key Summary GSV 150 Announcement : The Top 150 Global Companies in AI-Powered Digital Learning and Job Training Scale : 150 companies with combined annual revenue of $50 billion and over 3 billion cumulative learners Improving profitability : Two-thirds of companies are cash-flow positive, and EBITDA is expected to increase by 14% by 2025. High-growth companies : Cursor, Synthesia, MagicSchool, and others with annual revenues exceeding $50 million and growing more than twofold each year. Wayground (formerly Quizizz) : Selected as India-based company in the K-12 sector. Top 150 out of 3,000 companies worldwide GSV analyzed over 3,000 companies worldwide to identify the 150 fastest-growing companies in digital learning and job training. They were evaluated based on revenue size, growth rate, user base, global scalability, and profitability. This year, for the first time, the list includes publicly traded companies. These 150 companies combined have annual revenues of over $50 billion and a cumulative student base of over 3 billion. Profitability has also improved significantly, with two-thirds recording positive cash flow and EBITDA up 14% year-over-year. This year alone, they've raised over $5.5 billion in investment. Developer tools on the education company list? What's interesting is the sheer number of companies on the list. They include developer AI tool Cursor, AI video production Synthesia, language learning Speak, voice AI ElevenLabs, coding education Replit, and teacher AI MagicSchool. All of them generate over $50 million in annual revenue and are growing more than twofold each year. AI-powered audio and video-based learning experiences are blurring the boundaries of education. We live in an era where coding tools become learning tools, and video production tools become educational content. Where and for whom do you grow? By sector, K-12 (53%), career education (43%), and higher education (37%) dominate. Regionally, North America dominates with 64%, followed by Europe (14%), India (9%), and China (6%). What Wayground Proved Wayground was selected as an India-based K-12 company. With a philosophy of "Teachers, not AI," it has been creating tools that give teachers back time to be creative. This selection is proof that the market is accepting the idea that technology is a tool and people are at the center. https://www.asugsvsummit.com/gsv-150