Dr. Fridolin Ting is the co-founder of YoChatGPT!, a platform for collaborative learning with AI, and a Senior Lecturer at The Education University of Hong Kong. He conceptualizes and develops active teaching pedagogies that integrate generative AI to enhance critical thinking and teamwork in STEM education.
Editor’s Note:
The recent launch of ChatGPT’s group chat feature signals something significant: AI in education may be entering a new phase. After two years dominated by 1:1 AI tutors and individual productivity tools, we’re seeing the emergence of what might be called “Social AI”—systems designed not to replace human interaction, but to make it work better.
We’re excited about this shift. Tools like ChatGPT’s group chat, YoChatGPT!, Oko Labs, Breakout Learning, Human2Human, Class Technologies, and more are pioneering a different approach: AI that brings students together rather than pulling them apart, that facilitates peer collaboration rather than substituting for it, and that makes teachers more effective rather than redundant. If these early experiments prove successful, 2026 could mark the beginning of social AI as a strong force in education.
The article below makes the case for why this matters. It’s a vision we find compelling, and one we’ll be exploring more deeply in early 2026. Stay tuned!
AI in education as it currently stands conjures a solitary image: one student with one piece of technology getting personalized help or quick answers. This is a lonely, one-on-one AI experience. While born from the long held industry dream of achieving personalized learning, this paradigm is creating a generation of learners who relate to technology rather than through it. As Erin Mote and Michelle Culver powerfully argued in their Edtech Insiders article on Designing Student-Facing AI at the Speed of Trust, with this model we risk repeating the social media playbook: optimizing for engagement over well-being, and sacrificing human connection which is at the very bedrock of learning [1][2].
The industry is at a crossroads. We can continue down the path of isolated efficiency, creating ever-more-compelling digital companions that risk exacerbating student isolation. Or, we can choose a different path: one where AI is designed not as a private oracle for the individual, but as a catalyst for collaboration within a group. This is the promise of Social AI or Collaborative AI: tech designed to nurture the human relationships and creative problem-solving skills that form the foundation of learning. It prepares students for a future where collaboration—with both humans and AI—is paramount.
The Collaborative Pivot: Why Human-AI Teams Excel at Creation
The push for a more social, collaborative AI is not just a philosophical preference, it’s backed by scientific discovery! From Nobel Prize-winning breakthroughs in physics and chemistry, to AI-powered creative tools, we see a clear pattern: human-AI teams achieve incredible feats of creation. Emerging research confirms this. A pivotal 2024 meta-analysis by Vaccaro, Almaatouq, and Malone revealed a critical distinction: while human-AI teams often underperform human-only teams in pure decision-making tasks, they significantly excel in content creation. This includes everything from discovering move 37 in Go, to folding a protein behind heart disease, to generating eye-catching multimedia.
This insight is a game-changer for education. While we worry about students using AI to “decide” on a final answer for a test, the real opportunity lies in using AI to “create” possibilities. The messy, iterative, and divergent process of creative problem-solving—brainstorming, prototyping, debating ideas, and synthesizing diverse viewpoints—is precisely where collaborative AI can shine. Instead of an answers machine, AI becomes a creative partner, a Socratic guide, and a process facilitator. This aligns perfectly with UNESCO’s guidance for a “human-centered and pedagogically appropriate” approach, where AI supports, rather than replaces, human intellectual development.
This new “teacher-AI-student” dynamic reframes the goal from using AI to find the right answer to using AI to build the skills needed to generate great questions and create novel solutions as a team.
Designing for Collaboration: The Mechanics of Collaborative AI
Moving from Lonely AI to Collaborative AI requires a fundamental redesign of the user experience and the AI’s core function. At YoChatGPT!, we are building a platform based on this principle, which is part of a broader movement of “socially oriented” AI tools like BoodleBox, Oko Labs, and Breakout Learning who are all tackling this challenge. The architecture of this new model rests on three pillars:
1. AI as a Group Facilitator, Not a Lone Tutor
In a multi-user environment, AI’s role shifts from a knowledge dispenser to a process moderator. Its primary function is to enhance the group’s collaborative dynamics. For example, in a science class brainstorming solutions to a climate change problem, the AI can be prompted to:
Ensure Equitable Participation: “I’ve heard great ideas from Sarah and Ben. What are your thoughts, Maria?”
Introduce Constructive Friction: “That’s an interesting approach. What are the potential downsides or ethical implications we should consider?”
Synthesize and Scaffold: “The group has discussed solar and wind power. How might we combine these ideas into a hybrid solution?”
This approach uses prompt engineering techniques and pre-defined triggers within a group setting, to foster deeper critical thinking rather than surface-level responses. The AI doesn’t give the answer; it improves the conversation that leads to the answer.
Room in YoChatGPT! where 100+ participants contributed their thoughts on if generative AI has revolutionized education. The AI moderator helped categorize attitudes and biases, automatically, towards genAI, in real time.
2. Teacher-by-Design: Transparency Overcomes the Trust Deficit
A major concern with lonely AI is that the interaction is a “black box,” leaving educators to guess how students are using the tool. This fuels fears of misuse and academic dishonesty. Collaborative AI flips this dynamic. By bringing AI into a collaborative, observable space, platforms like YoChatGPT! provide a dashboard for educators. Teachers can see the entire interaction transcript, monitor group progress in real-time, and identify not only how an individual user learns with GenAI, but also which teams are thriving and which are stuck.
This transparency is the key to building guardrails and trust when integrating GenAI into K-12 or higher education classrooms, where teacher oversight is not a bug—it is an essential feature for safety, accountability, and effective pedagogical intervention. It allows the teacher to move from being a “sage on the stage” to a “guide on the side,” armed with real-time data on student learning with generative AI and group dynamics.
3. Building Future-Ready Skills
The ultimate goal of education is not just knowledge acquisition but skill development. By working in an AI-facilitated group, students are not just learning course content; they are building essential, future-ready competencies:
AI Literacy: They learn to treat AI as a tool to be managed and directed within a team, a crucial skill for the modern workforce.
Collaborative Problem-Solving: They practice negotiation, debate, and synthesis with their human peers, with the AI acting as a scaffold.
Critical Thinking: They are prompted to question assumptions, consider alternative viewpoints, and defend their reasoning—both to their peers and to the AI.
The edtech industry has a choice. We can continue to build tools that, however well-intentioned, pull students away from each other. Or we can build tools that bring them together in more powerful ways. The path of Collaborative AI is not about slowing down innovation, it’s about directing that innovation toward a more human-centric and pedagogically sound future.
This requires a shared commitment from developers, investors, and educators. We must measure success not by time-on-app, but by the quality of the collaboration, the depth of the learning, and the creation of novel solutions to old problems. We must build trust through transparency and design AI tools that strengthen the human relationships forming the infrastructure of learning and work.
This is more than a product decision; as the “Speed of Trust” article noted, it’s a social contract. We invite our peers and partners across the edtech landscape to join us in this work. Let’s co-design pilot projects, share our findings, and build a future where AI’s greatest contribution to education is not the answers it provides, but the connections it helps us build.
[1] Bücker, S., Nuraydin, S., Simonsmeier, B. A., Schneider, M., & Luhmann, M. (2018). Subjective well-being and academic achievement: A meta-analysis. Journal of Research in Personality, 74, 83–94. https://doi.org/10.1016/j.jrp.2018.02.007
[2] Allen, K. A., Kern, M. L., Rozek, C. S., McInerney, D. M., & Slavich, G. M. (2021). Belonging: A review of conceptual issues, an integrative framework, and directions for future research. Australian Journal of Psychology, 73(1), 87–102. https://doi.org/10.1080/00049530.2021.1883409
We love to collaborate. To learn more about partnership and sponsorship opportunities, please email info@edtechinsiders.com. Thanks for reading!
Top Edtech Headlines
1. College Students Flock to a New Major: AI
Students at top U.S. universities are rapidly pivoting from traditional computer-science tracks to specialized AI majors. At MIT, the new “Artificial Intelligence and Decision-Making” program has already surged to the school’s second-most-popular major — signaling a reshaping of what the next generation wants from tech education.
2. Sam Altman’s ‘Code Red’ Memo Urges ChatGPT Improvements Amid Growing Google Threat, Reports Say
Sam Altman has reportedly issued a “code red” memo at OpenAI, ordering teams to double down on improvements to ChatGPT—focusing on speed, reliability, and personalization—as the company faces growing pressure from Google Gemini 3 and other rivals.
3. OpenAI Develops New AI Model Called ‘Garlic’ to Compete Against Alphabet’s (GOOGL) Gemini 3
OpenAI is reportedly building a new AI model codenamed Garlic — a fast-tracked next-gen model aimed at challenging Gemini 3 (from Alphabet) and Opus 4.5 (from Anthropic). Early internal tests suggest Garlic is outperforming those rivals in coding and reasoning tasks, with a potential public release as GPT‑5.2 or GPT‑5.5 slated for early 2026.
4. India’s First Consumer Edtech IPO: Celebrating PhysicsWallah’s Public Market Debut
PhysicsWallah — once a scrappy YouTube tutoring channel — has officially become India’s first edtech company to debut on the public markets. Its shares surged more than 30% on listing day, underscoring renewed investor confidence in edtech just as the sector attempts a rebound.
This edition of the Edtech Insiders Newsletter is sponsored byTuck Advisors.
Since 2017, Tuck Advisors has been a trusted name in Education M&A. Run by serial entrepreneurs with over 30 years of experience in founding, investing in, and selling companies, you deserve M&A advisors who work as hard as you do.
Not all LOIs are the same. Use our free Deal Value Calculator to estimate the relative expected value of your LOIs.
Charlie is the co-founder of Trophy.so, a gamification infrastructure platform helping edtech and wellness companies build engaging learning experiences in weeks, not months. With a background in product development and motivation design, Charlie focuses on creating scalable systems that boost retention, habit formation, and learner satisfaction.
