What it is
AutCraft is a whitelisted Minecraft server created specifically for autistic players and their families. It was founded in November 2013 by Stuart Duncan, a web developer in Timmins, Ontario, whose eldest son is autistic. Duncan, known in-game as AutismFather, expected perhaps a dozen children to sign up. Within two days, 750 people had tried to join. As of early 2026, the server has over 20,000 whitelisted players.
The server runs a modified version of Minecraft with accommodations built into the infrastructure. These include chat modifications for players who find text communication difficult, an augmentative and alternative communication (AAC) system for players who are non-speaking or who prefer not to use the main chat, modified social rules that prioritise support over competition, and a reward structure that recognises positive community behaviour rather than combat skill or resource accumulation. Several of these accommodations would be considered “cheating” on a standard Minecraft server. On AutCraft, they are design choices that reflect the community’s understanding of what its players need.
The server is administered by a team of neurodivergent and neurotypical adults, with support from community members. It operates as a safe space in the specific sense that bullying, griefing (destroying other players’ creations), and discriminatory language result in consequences, and that the social norms are explicitly communicated rather than left implicit. For autistic players who have been excluded from or bullied in other online gaming spaces, this is not a minor feature. It is the reason the server exists.
Why it matters
AutCraft is significant for several reasons, and they are not all the ones that media coverage tends to emphasise.
The media story, repeated across ABC News, CBC, NPR, PC Gamer, and a 2017 TEDx talk by Duncan himself, is heartwarming and true as far as it goes: a father built a safe online world for his autistic son and thousands of other children. That story matters. But it is not the whole of what AutCraft demonstrates.
It is autistic community built on autistic terms. The server’s social norms were not designed by therapists or clinicians. They were developed by and with the community. The modifications to standard Minecraft gameplay are not therapeutic interventions. They are accommodations that the community identified as necessary for its members to participate fully. This is the social model of disability (see Models of disability) applied to a virtual environment: the barrier to participation was not in the players but in the default design of the game, and changing the design removed the barrier.
It challenges the social deficit narrative. One of the most persistent claims about autism is that autistic people are not interested in social interaction. AutCraft is a social space. Its players form friendships, collaborate on building projects, celebrate events together, support each other through difficulties, and develop a shared culture with its own vocabulary. The community coined the word “autsome” (a portmanteau of “autism” and “awesome”) to describe the positive identity they share. This is not socialisation as a neurotypical would define it: it happens through text rather than speech, through shared building rather than face-to-face conversation, on players’ own schedules rather than in real-time synchronous interaction. But it is social, and the claim that autistic people lack social motivation looks different when 20,000 of them have joined a community whose entire purpose is social connection.
This aligns with what the double empathy problem predicts (see The double empathy problem). Autistic people do not have a deficit in social motivation. They have a mismatch with the social environments designed by and for neurotypical people. Provide an environment where the social norms are explicit, the communication channels are accessible, and the sensory demands are manageable, and autistic sociality flourishes. AutCraft is an existence proof.
It generated serious academic research. Kathryn Ringland, then at UC Irvine, conducted an extended ethnographic study of AutCraft that produced some of the most detailed research available on autistic social behaviour in online spaces. Her work, spanning from 2015 to 2022, is unusual in the autism research landscape because it studies autistic people doing something they chose to do, in a space they built for themselves, rather than measuring their performance on tasks designed by non-autistic researchers.
The research
Ringland’s body of work on AutCraft is worth knowing in detail because it represents participatory, community-embedded research of the kind the wiki’s epistemic justice pages advocate (see Participatory research and epistemic justice).
Ringland et al. (2015) published “Making ‘Safe’: Community-Centered Practices in a Virtual World Dedicated to Children with Autism” at CSCW (Computer-Supported Cooperative Work), documenting how the community developed its own safety practices and social norms.
Wolf, Faucett, Dombrowski, Hayes, and Ringland (2016) published “‘Will I Always Be Not Social?’: Re-Conceptualizing Sociality in the Context of a Minecraft Community for Autism” at CHI (the premier human-computer interaction conference). The title captures the paper’s argument: the standard framing of autistic people as “not social” fails to account for the rich social interaction observable in spaces designed for autistic sociality.
Ringland, Wolf, Boyd, Baldwin, and Hayes (2016) published “Would You Be Mine: Appropriating Minecraft as an Assistive Technology for Youth with Autism” at ASSETS (the ACM conference on assistive technology). The paper documented how AutCraft members used Minecraft’s building tools to create their own assistive technologies within the game: communication aids, sensory regulation spaces, social scripts embedded in the environment. This paper won Best Paper at ASSETS 2016, and its central finding is important: the community did not wait for researchers or clinicians to design assistive tools for them. They built their own, using the materials available, in response to needs they understood better than anyone outside the community could.
Ringland (2019) published “‘Autsome’: Fostering an Autistic Identity in an Online Minecraft Community for Youth with Autism” at iConference, examining how the community’s shared vocabulary and practices contributed to positive autistic identity formation. The word “autsome,” the paper argues, is not just a playful coinage. It is a community reclaiming the language around autism, turning a label that has been used to marginalise into one that carries pride. For children who encounter negative messaging about autism in school, in clinical settings, and online, having a community that treats autism as something to celebrate rather than overcome has developmental significance.
Ringland and Wolf (2022) published a reflection piece in ACM SIGACCESS Accessibility and Computing looking back at their work five years on. They argued that their most important finding was methodological: engaging with the community and letting its members lead the conversation had allowed them to see assistive technologies “in a broader frame,” technologies the community was crafting for itself rather than receiving from outside experts. This is participatory research in practice, not just principle.
What it is not
AutCraft is not a therapy programme. It is not an intervention. It was not designed to teach social skills or to reduce autistic behaviours. Duncan has been clear about this: the server exists so that autistic children can play a game they love in a community that accepts them. That the experience also appears to support social development, positive identity, and community belonging is a consequence of good design, not a clinical goal.
This distinction matters because AutCraft’s success is sometimes co-opted by narratives about “using Minecraft to treat autism.” The server is not treating anything. It is providing an environment in which autistic people can be themselves, and it turns out that when you do that, good things follow. The lesson is about environment design, not about intervention (see Epistemic justice and neurodivergence).
Key sources
- Duncan, S. (2017). How I use Minecraft to help kids with autism. TEDxYorkU. Available at: https://www.ted.com/talks/stuart_duncan_how_i_use_minecraft_to_help_kids_with_autism
- Ringland, K.E., Wolf, C.T., Dombrowski, L. and Hayes, G.R. (2015). Making ‘safe’: community-centered practices in a virtual world dedicated to children with autism. Proceedings of the 18th ACM Conference on Computer Supported Cooperative Work and Social Computing (CSCW ‘15), 1788-1800. doi: 10.1145/2675133.2675216
- Wolf, C.T., Faucett, H., Dombrowski, L., Hayes, G.R. and Ringland, K.E. (2016). ‘Will I always be not social?’: re-conceptualizing sociality in the context of a Minecraft community for autism. Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems, 1256-1267. doi: 10.1145/2858036.2858038
- Ringland, K.E., Wolf, C.T., Boyd, L., Baldwin, M. and Hayes, G.R. (2016). Would you be mine: appropriating Minecraft as an assistive technology for youth with autism. Proceedings of the 18th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS ‘16), 33-41. doi: 10.1145/2982142.2982172
- Ringland, K.E. (2019). ‘Autsome’: fostering an autistic identity in an online Minecraft community for youth with autism. Information in Contemporary Society: iConference 2019 Proceedings, Lecture Notes in Computer Science 11420, 132-143. doi: 10.1007/978-3-030-15742-5_12
- Ringland, K.E. and Wolf, C.T. (2022). Creating assistive technology in disabled communities, five years on: a reflection of neurodivergency and crafting accessible social spaces. ACM SIGACCESS Accessibility and Computing, 131. doi: 10.1145/3507912.3507914