THE FIFTH PARAMETER: Issue 004: The Adaptation Burden Why Accessibility Still Requires People to Adapt to Systems

The accessibility sector has become remarkably effective at measuring absence. We can quantify caption accuracy, calculate speech recognition error rates, audit websites for compliance failures, and produce increasingly sophisticated dashboards that catalogue where barriers remain. Yet there is a curious asymmetry at the center of contemporary accessibility practice. We have become proficient at measuring defects within systems while remaining comparatively incapable of measuring whether systems are designed around the people expected to use them. Accessibility is still too often conceived as a corrective measure applied after the fact rather than as a governing principle embedded at the outset.

This distinction matters because the emerging generation of AI-enabled accessibility technologies is frequently presented as evidence of progress. New tools appear almost weekly: automated captions, speech-to-text systems, text-to-speech applications, synthetic voices, sign language avatars, and machine translation platforms. Collectively, they suggest a future in which barriers can be reduced through technical innovation alone. Yet beneath this optimism lies a more fundamental question. Why do individuals continue to bear the burden of adapting themselves to institutions? Why is accessibility still largely organized around helping people navigate systems rather than requiring systems to recognize and accommodate people from the moment interaction begins?

The prevailing model of accessibility assumes a recurring negotiation. A Deaf individual must request an interpreter. A DeafBlind individual must explain communication needs. A person requiring alternative formats must identify themselves repeatedly across agencies, schools, healthcare providers, employers, and digital platforms. Each interaction begins with disclosure, explanation, and accommodation. Technology may reduce some friction within this process, but it rarely challenges the underlying architecture that makes such negotiations necessary. The burden remains with the individual. The system remains unchanged.

What emerges from this arrangement is a paradox. The more sophisticated accessibility technologies become, the easier it becomes to overlook the structural conditions that necessitate them. AI can translate, transcribe, summarize, and generate. Yet these capabilities address symptoms more readily than causes. A system that instantly produces captions still assumes that users must enter an environment designed without them in mind. A sign language avatar may deliver information more effectively, but it does not answer why sign language access was absent from the original design process. Technical capability can obscure institutional responsibility.

This raises the possibility that the next major development in accessibility will not be another tool at all. It may instead be a new layer of governance operating beneath individual technologies. Imagine a communications infrastructure in which preferences are recognized before interaction occurs. A citizen’s preferred language, communication modality, accessibility requirements, and accommodation needs would not require repeated disclosure. Rather than requesting access, individuals would encounter systems capable of adapting automatically and appropriately. Accessibility would cease to function as an exception process and would instead become a characteristic of the environment itself.

Such a vision is often framed as a technological challenge. In reality, it is primarily a governance challenge. The central questions are not about engineering capability. They concern stewardship, accountability, and trust. Who maintains communication preference data? Who determines standards? Who validates accuracy? Who ensures interoperability across institutions? Who is liable when a system fails to recognize a person’s needs? Most importantly, who ensures that such infrastructures serve the interests of citizens rather than the interests of vendors seeking to monetize accessibility data?

These questions become particularly urgent as governments, universities, healthcare systems, and private organizations accelerate their adoption of AI-mediated accessibility services. Discussions frequently focus on what these systems can do. Far less attention is devoted to how they should be governed. Accessibility has historically been treated as a matter of compliance. Artificial intelligence transforms it into a matter of public stewardship. Decisions regarding procurement, oversight, transparency, accountability, and risk management increasingly determine whether accessibility technologies empower communities or merely create new forms of dependency.

The accessibility community therefore faces a choice. One path continues the familiar cycle of identifying barriers and developing increasingly sophisticated tools to help people navigate them. The other asks a more ambitious question: what would it mean to build institutions that adapt to human diversity by default? The first path is technological. The second is political, organizational, and ethical. It concerns the distribution of responsibility between individuals and systems.

Perhaps the future of accessibility will not be defined by better avatars, more accurate transcription engines, or faster translation models. Perhaps it will be defined by something less visible but ultimately more consequential: the emergence of governance frameworks that make adaptation an obligation of institutions rather than a burden placed upon citizens. The most important accessibility innovation of the coming decade may not be another tool. It may be the decision that people should no longer have to ask to be accommodated at all.