Pro-Social AI in practice: how Da Vinci Schools is using technology to strengthen human support systems. At Da Vinci Schools, a growing network of mental health and wellness partnerships ensures students get the support they need - but connecting each student to the right resources at the right time can be complex. Now, in an innovative next step, Da Vinci is using pro-social artificial intelligence to help counselors and staff navigate these needs and deliver more personalized support - transforming how student care happens behind the scenes. Since its inception, Da Vinci Schools has prioritized building a comprehensive student support ecosystem rooted in health and wellness partnerships - ensuring every young person has access to the care they need to fully engage in school and life. Under the leadership of Assistant Superintendent Dr. Erin Whalen, the network has expanded on the groundbreaking work of Da Vinci RISE High School, developing systems and partnerships that support students across a wide range of needs. In close collaboration with leaders such as Dr. Emily Green, Dr. Lynn Rodriguez, and a dedicated team of counselors and psychologists, Da Vinci has cultivated an expansive, coordinated network of providers - including Hazel Health, Clear Behavioral Health, allcove, STAR Wrap Service, Care Solace, Better Youth, and others. As this network of support grew, so did a critical challenge: knowing when, where, and how to connect students to the right resources at the right time. The solution came through an unexpected and innovative approach - leveraging pro-social artificial intelligence. Recognizing the opportunity, Whalen partnered with Dr. Timothy Regalado, a highly respected AI developer and practitioner, to build a secure internal AI app designed specifically for Da Vinci's student services teams. The system draws from an extensive body of work developed over the past year and a half - board policies, counseling protocols, referral guides, and operational documents previously housed within the network's Basecamp system - helping staff navigate complex needs and deliver timely, personalized support to students. These materials were transformed into a closed, secure AI companion that helps counselors, administrators, and school psychologists navigate complex student cases and determine appropriate referral pathways. This tool is powered by, an educator friendly platform for creating custom AI tools. Playlab has partnered with Da Vinci Schools for the past two years, helping staff and students build AI literacy and develop tools to support teaching and learning, including this one. Unlike many public AI systems, such as ChatGPT or Claude, which may store or train on potentially sensitive user data, Da Vinci's tool operates within strict safeguards: * *All student information remains anonymous* * The system is closed-source and references only internal documents * No external data is accessed, reducing the chances of errors or data leakage Furthermore, rather than replacing human decision-making, the AI functions as a "thought partner" - guiding staff through protocols, suggesting appropriate resources, and reinforcing consistent practices across the organization. Crucially, the system is designed to be pro-social. It intentionally directs human professionals back to one another and reinforces Da Vinci's collaborative leadership structure. Housing instability cases, for example, are routed to Erin Whalen as Da Vinci's McKinney-Vento Liaison. Special education matters connect to Lynn Rodriguez, Executive Director of Special Education, while health and wellness issues are directed toward Emily Green, Health Services Director. In this way, the technology strengthens - rather than replaces - the human relationships at the center of the system. This approach reflects a broader philosophy emerging at Da Vinci Schools: AI should not distance people from one another, but deepen its capacity for care, clarity, and coordinated action. At its best, this form of pro-social AI does exactly that - empowering educators to act with greater confidence, ensuring consistency across complex systems, and ultimately helping schools respond more effectively to the real challenges facing young people and families.

Enabling more people to shape AI might transform its education system. When educators create with AI, they realize it isn't magic, it's a new, malleable design material. One that they can adapt to meet their needs and use to bring their visions of education to life. This week, Playlab Education, Inc. had the chance to demo Playlab at Google.org's AI for Learning Forum. This is what Playlab Education, Inc. shared during its demo. Playlab sits at the intersection of three insights: First, Playlab Education, Inc. know its education systems need to change. This was true before AI. AI makes this change more urgent. Second, AI makes it possible for more people to participate in changing its education systems. How? In the coming years, anyone will be able to create software. Playlab makes this easier to do in education. Third, Playlab Education, Inc. don't yet know how to harness AI in education within its current systems or how it might enable Playlab Education, Inc. to create better education models. Rapidly scaling solutions now feels unwise. Rather, it seems more apt to create room for divergent experimentation - to learn and discover the myriad ways in which Playlab Education, Inc. might use AI to improve and transform education. Just like the advent of video cameras required intrepid writers and filmmakers to invent cinema, Playlab Education, Inc. need forward-looking educators to help invent the myriad ways in which AI might enable improvements in teaching and learning. Today, its community creates chat-based apps grounded in their knowledge and values. In the near future, they'll build all kinds of educational software. Its aim is to create construction kits so anyone can build robust AI edtech. On a fully open stack - open applications, open software and open AI infrastructure. Because AI systems need to be transparent, legible and malleable. For example, Building 21 is a network of project-based schools. Students don't get traditional grades. Instead, they are evaluated on competencies that they develop over time through real world projects. The challenge? Designing authentic projects is hard. Coaching and supporting students on those projects is even harder. Especially with the constraints of public education, which means this work often varies in quality. Using Playlab, Building 21 has built a collection of AI apps for students and teachers. Let's jump into this example: It's called "Parker," and it's a competency-based feedback tool. A student drops in a piece of work, and based on what they're working on, gets specific, competency-aligned feedback. Through their use of AI, they were able to move from teacher-designed projects to student-designed ones, all while deepening rigor. So how did they do it? Let's peek inside. But let's look at it through the perspective of a district in North Dakota whose educators found this app on Playlab's community site and "remixed it" to suit their needs, their context. Playlab Education, Inc. can start with the system prompt. You can see which models they tried - some of their schools are powered by Gemini 2.5 flash. By adding "References," they can add their own knowledge to the app, which means they can swap out Building 21's competencies for their own and add any additional knowledge they'd like the AI to harness. They can also connect this app to a range of Model Context Protocol (MCP)-based tools that ground it in deterministic outputs, something especially helpful in areas like math. As they build, educators get a hands-on feel for why AI isn't magic. It's a new design material. One that they can tinker with and adapt to meet their needs. As they build, they inform the next set of building blocks Playlab Education, Inc. at Playlab create - based on their goals, asks and feedback. For now they're building chat-based apps. In the coming months, they'll create interactive voice-based experiences, writing feedback tools, and AI-assisted writing tools. And in the coming years, they'll be able to create all kinds of AI-native interfaces for teaching and learning. Their ideas also spread. Because many choose to openly share their AI projects, other school systems can borrow and "Remix" them. Islands of innovation turn into networks of educators and systems working together to bring more ambitious learning to life in more places. Building 21 is just one example. Organizations like Leading Educators and Bank Street Education Center work closely with schools across the US to help teachers more effectively teach with high quality instruction materials (HQIM). Their work involves intensive coaching. Cycles of data analysis, classroom observations and feedback. To aid in this work, Leading Educators and Bank Street have built collections of AI applications that support coaches, leaders and teachers. For example, Leading Educators is working closely with school systems like Boston Public Schools and KIPP North Carolina to help both systems implement Illustrative Mathematics (IM). Because real change requires working up and down a system, Leading Educators has created AI applications that support their work with teachers and leaders. This IM Lesson Coach helps teachers break down, understand and teach with IM. It's not meant to replace coaching, but is designed specifically as a supplement. The app is grounded in Learning Commons' KnowledgeGraph that includes IM, learning components from ANet, and standards. It's also grounded in Leading Educators' frameworks on coaching. As teachers use their IM Lesson Coach, Leading Educators can easily iterate on the app based on user feedback and real world use. Because Leading Educators is committed to open sharing, other nonprofits like Bank Street or districts like Boston can borrow, learn from, and adapt their work. Both examples offer two different ways in which teams with deep expertise - in competency based learning and in HQIM implementation - are using Playlab to create AI applications that reflect their specific needs, workflows, and knowledge. They built on open infrastructure instead of starting from scratch and can learn with and from a global community of educators creating with AI. The Playlab team comes from the world of end-user programming and creative computing. Playlab Education, Inc. is deeply committed to open technology because Playlab Education, Inc. think the systems Playlab Education, Inc. is building and using in education need to be technically transparent, legible and malleable so that people can make them their own. Its hope is to create a world where many more people can create software that helps them grow their capabilities. Along the way, they'll help Playlab Education, Inc. grow and evolve to more humane, empowering education systems that better serve all learners. Thanks to support from Google.org, Playlab Education, Inc.'ll be able to bring this work to more communities starting in Idaho, Indiana and Tennessee.