The Case for Social AI in Education
Why the Future of EdTech Might Be More Collaborative Than We Think
Thanks to our Presenting Sponsors Hire Education, Starbridge, Cooley, and Tuck Advisors for making Edtech Insiders’ content and community possible.
The Case for Social AI in Education
Why the Future of EdTech Might Be More Collaborative Than We Think
By: Alex Sarlin
Meta’s $2 billion acquisition of Manus in the final days of 2025 put AI agents squarely at the center of conversation, and autonomous agents are likely to be a major trend in educational AI over the next year. But these AI agents remain fundamentally solitary tools. One person, working in isolation with one (or more) AI agents.
Amid the fanfare around AI agents, an overlooked product launch from late 2024 points toward a very different future, one that I find much more intriguing. In November, OpenAI quietly rolled out Group Chats in ChatGPT—allowing multiple people to collaborate with each other and with ChatGPT in the same conversation.
Unlike the solo AI agent paradigm dominating headlines, multi-user collaborative AI represents something fundamentally different and truly exciting: AI as social infrastructure that strengthens relationships and encourages collaboration. This is why one of my predictions for 2026 is that it will be the year that “Social AI” breaks into the mainstream, especially in education.
What Is Social AI?
In a recent article we published by Dr. Fridolin Ting, co-founder of Higher Education Social AI startup YoChatGPT!, he captured the current state perfectly:
“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... creating a generation of learners who relate to technology rather than through it.”
Since ChatGPT’s 2022 launch, we’ve lived in the “chatbot era” of AI—most people’s experience with AI is primarily text-based, conversational, and one-on-one. Even tools that generate images, music, videos, or shareable coding artifacts tend to follow this solo paradigm. You use them alone, then share the output with other people.
But as Dr. Ting argues, there’s a different path possible: “AI 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.”
AI Facilitation
While research on Social AI in Education is quite nascent, there are some interesting possibilities raised by studies of AI facilitation among adults. Google Deepmind’s Habermas Machine (Science, 2024), demonstrated that AI facilitation could help diverse groups find common ground on divisive political issues, while a study on “Relational AI“ showed that relational AI facilitation (vs. personalized AI assistance focused on each participant individually) improved cooperation rates among people with different political opinions.
Why Social AI Matters for Education
Research has long demonstrated the positive impact of strong teacher-student relationships on student achievement, motivation, and self-efficacy. Social approaches to learning that emphasize student relationships, such as project-based learning, collaborative learning, and peer tutoring, have shown strong positive effects as well.
The main roadblock to getting to these social learning experiences is often logistics. Managing collaborative learning at scale is exhausting for teachers, as is maintaining strong individual relationships with large groups of students. Who works with whom? How do you ensure productive discussions? How do you support every group when you can only be in one place?
In these contexts, AI could play a strong role in:
Facilitating Group Dynamics
Acting as a mediator or facilitator, helping students navigate disagreements, and ensure all voices are heard. Or, helping teachers manage relationships with many students at once. AI can also follow particular pedagogical principles of group work: assigning roles to students, forming heterogeneous groups, supporting the big questions and artifacts of project-based learning, or ensuring equal participation for female students in math and STEM discussions.
Matchmaking Learners
Pairing students based on complementary strengths, similar struggles, or aligned interests. Or, pairing students with tutors or teachers likely to build strong relationships
Surfacing Hidden Connections
Noticing that two students are wrestling with the same misconception and bringing them together—not to get answers from each other, but to realize they’re not alone.
Providing Relational Support
Finding unlikely relational connections and overlaps to strengthen learning experiences: “Erica loves ice skating and Jim plays piano—why don’t you both talk about the role of practice in refining a skill?” That’s not a pedagogical intervention, it’s a relationship intervention.
Reflecting Patterns
As evidenced by the political studies, AI can synthesize, mediate, and help students find common ground, even in areas of disagreement: “It sounds like you’re all trying to solve for fairness, but defining it differently.”
Social AI Addresses Core AI Concerns
In an age of impending AI backlash, Social AI directly addresses four major anxieties driving resistance to AI in schools:
Screen Time and Isolation
Social AI should be designed to facilitate face-to-face, synchronous or asynchronous peer interaction. The AI’s job is to get students talking to each other, not to the screen, and the best tools will be the ones that feel the most lively, social and connected.
Replacing Teachers
Social AI can supports teachers in focusing on irreplaceable human work—building relationships, providing mentorship—while handling the logistics of small-group management and communication that currently make collaborative learning hard to scale.
Students Outsourcing Thinking
When AI facilitates discussion rather than answering questions, students should still be expected to do the cognitive work. Social AI, unlike many chatbots, isn’t giving answers; it’s prompting students to articulate reasoning and build on each other’s ideas.
Lack of Transparency
Educators struggle to differentiate meaningful student work from AI-generated answers, or to know how students are using AI in their thought processes. Social AI is, by necessity, transparent and auditable, allowing the process of building and collaborating to be visible to teachers.
Who’s Building It?
Luckily, a number of startups (and, as always a couple of AI giants) are starting to move into the Social AI space. These are some to watch in the next year:
Breakout Learning specializes in AI-facilitated peer-to-peer learning for professional development and higher education. Their platform creates structured collaborative experiences where AI guides learners through problem-solving exercises together, ensuring everyone contributes while keeping discussions focused and productive.
Disco supports cohort-based courses, community spaces, and AI automation in one platform— using AI to foster connection between community members, make matches, and spur conversation. Disco helps training businesses, customer academies, and enterprise L&D teams deliver learning experiences where people learn from each other in addition to content.
Honor Education exemplifies asynchronous social AI through “social annotation” that turns course materials into living documents. As students read, they can highlight passages as “unclear” or “important”—reactions that remain visible to future cohorts— and see notes left from other students—while instructors can embed video commentary directly in the text, creating what Honor calls “networks of practice” where each group of learners benefits from the questions and insights of those who came before them.
Human2Human AI positions AI as a bridge-builder between learners. Rather than promoting AI as the primary teacher, its platform uses AI to help students discover commonalities, surface complementary skills, and build the kind of peer relationships that support long-term learning communities.
OKO Labs focuses on making small group work manageable at scale by using AI to help teachers form effective groups, monitor discussions across multiple breakout rooms simultaneously, and provide targeted interventions when groups get stuck. Founder Matt Miller’s insight—that “too many classrooms, especially in underserved communities, don’t have the extra teaching assistants to make small group work happen effectively”—drives their mission to democratize access to collaborative learning.
PeerTeach supports peer tutoring through AI facilitation and just-in-time training, allowing students to follow the best evidence-based practices of successful peer tutoring and build relationships while supporting one another academically.
Swivl’s M2 Intelligent Teaching Assistant is designed to facilitate groupwork with an in-class AI robot that can support group dynamics
YoChatGPT, co-founded by Dr. Ting, is an AI-powered platform designed for collaborative learning where the AI moderates discussions instead of providing direct answers. This focus on collaboration has demonstrated a significant impact on students’ critical thinking and academic performance.
Last— but never least— OpenAI’s experiments with Group Chats hint at a sliding door future in which the next version of ChatGPT is social (SlackGPT?). Similarly, Google‘s world-class infrastructure for real-time collaboration in Docs and Workspace makes it a natural candidate to build social AI for education; adding AI that could actively facilitate group learning would be a logical extension of tools already embedded in schools worldwide.
Top Edtech Headlines
1. Amazon Reiterates AI Literacy Focus Following White House Education Meeting
Amazon emphasized its commitment to expanding AI literacy and workforce development after senior AWS leaders met with U.S. Secretary of Education Linda McMahon, highlighting support for national efforts to embed AI education across K–12, higher ed, and workforce training. The company reinforced its ongoing involvement in initiatives like the White House’s Pledge to America’s Youth, stressing upskilling and reskilling programs for students, educators, and workers to prepare them with essential AI skills.
2. Open AI’s Upcoming AI Gadget to Be Controlled by Voice
OpenAI is developing a new AI hardware device equipped with cameras, microphones, and speakers that will be primarily voice-controlled since it won’t include a screen. The company is working on a more natural conversational voice model for the gadget, which is expected to launch no earlier than fall 2026 once those systems are ready.
3. A New Podcast Examines the Crisis in Education—and Paths Forward
Dr. Charles Reigeluth, Professor Emeritus at Indiana University, has released a new 10-chapter podcast series exploring why American education needs a deeper paradigm shift. The Crisis in Education: What It Is and How to Fix It examines personalized, competency-based, and project-based learning through practical examples, case studies, and concrete ideas for transforming schools and systems.
4. Teens Favor “Best Friend” AI Chatbots, Raising New Safety Questions for Edtech
New research from the University of Denver shows that adolescents overwhelmingly prefer “best friend”-style AI chatbots that use relational, human-like language over transparent systems that clearly state they aren’t human, raising safety and design concerns for EdTech tools aimed at young users. The findings suggest that conversational tone strongly influences emotional trust and engagement, especially among socially or emotionally vulnerable teens, which could heighten risks without strong AI literacy and boundary cues in educational technology.
What If School Worked Like a Video Game?
We recently had Noah Bushnell and Dr. Leah Hanes on The Edtech Insiders Podcast!
Nolan Bushnell is the founder of Atari and Chuck E. Cheese and is widely known as the father of the video game industry. Dr. Leah Hanes is the CEO of ExoDexa, co-author of the ExoDexa Manifesto, and leads the Two Bit Circus Foundation, where she has impacted over 460,000 students through game-based and project-based learning.
5 Things You’ll Learn in This Episode:
Why traditional schooling fails students who don’t learn at the “right” pace
How game design principles like flow and mastery can transform learning
The role AI can play in personalized, no-grades education
Why gamification efforts led by academics often fall short
What a “perfect school day” could look like in an AI-driven future
We love to collaborate. To get in touch, email info@edtechinsiders.com. Thanks for reading!