Decoding the Unsaid: Social Opacity, Autism, and AI as Accessibility Infrastructure

This conceptual article proposes an environmental framework for understanding social difficulty in autistic adulthood, one that complements rather than replaces existing cognitive accounts. Many autistic adults face employment instability, health care avoidance, relationship rupture, and chronic burnout; predominant frameworks have located these difficulties within the individual in cognitive or communicative differences understood to reduce social success. I argue that a meaningful portion of this difficulty may instead arise from environments that depend structurally on hidden, unstated norms—a condition I term social opacity. I introduce the norm accessibility gap, decomposed into two independently variable environmental dimensions—norm implicitness and explicit scaffolding—in a way that identifies concrete targets for intervention. Drawing on the social model of disability, the double empathy problem, and neurodiversity scholarship, I argue that social difficulty is in part relational and ecological. Against this backdrop, I propose that Artificial Intelligence (AI) systems may eventually function as social accessibility infrastructure—carefully distinguished from individual assistive tools, with significant implications for institutional responsibility and policy. I distinguish two operational modes (interpretive and mediating) corresponding to the two dimensions of the norm accessibility gap and organize applications across a three-stage developmental pathway from individual tools toward institutional infrastructure. The goal of such tools is not normative imitation but increased autonomy, reduced cognitive burden, and reciprocal understanding—not making autistic adults more neurotypical, but making their social environments more accessible. I attend to ethical risks and propose a participatory research agenda led by autistic adults across diverse support profiles. Community Brief This section summarizes the article in plain English for autistic adults, families, clinicians, advocates, and community readers. This article is conceptual—it does not report a clinical trial or empirical study. What is this article about? This article looks at why social life can be so difficult to navigate for many autistic adults—and asks whether AI technology could eventually help. Rather than focusing on what autistic people might find challenging, it examines the social environment itself. Many social settings depend on unwritten, unstated rules that nobody explains. The article calls this social opacity. It introduces a new concept—the norm accessibility gap—to describe how much of a given setting’s social rules are hidden and unexplained, and asks whether AI tools could help reduce this gap. Why is this topic important? Autistic adults face real and serious consequences when social environments rely on hidden norms: job loss, health care avoidance, relationship difficulties, and exhaustion. Most research has focused on helping autistic people adapt to these environments. This article argues that approach places the entire burden on autistic adults—and that environments themselves could be designed more accessibly. This shift from a deficit lens to an environmental one reflects values central to the neurodiversity movement and the social model of disability. What are the key messages? Three ideas are central. First, social environments vary in how opaque they are—some make their expectations clear, others do not. Second, AI systems could potentially support accessibility in two ways: helping autistic adults interpret ambiguous social situations (an interpretive mode) or introducing explicit communication structure that benefits everyone involved (a mediating mode). Third, responsibility for social accessibility should rest with institutions, not solely with autistic adults. The article frames AI as potential infrastructure—like captioning—rather than just an individual assistive tool and describes a pathway in which near-term individual tools help build toward long-term institutional change. How could this help autistic adults? If these ideas are developed responsibly, autistic adults might eventually have better support navigating workplace communication, health care appointments, or situations where unwritten rules feel inaccessible. The goal is not to make autistic adults appear more neurotypical. It is to reduce unnecessary informational barriers in environments that could be made more transparent for everyone. What should happen next? This article is conceptual—it does not report data or test any intervention. The essential next step is participatory research led by autistic adults, including those with diverse communication styles and support needs. Autistic coresearchers should determine which questions matter, whether tools like these would actually be useful, and whether the framework accurately reflects autistic experience. No tool based on these ideas should be developed without that genuine co-leadership.