When AI Access Alliance talks about "AI comprehension tools," we mean something specific. Not all AI is the same. The tools we advocate for serve one purpose: helping people understand information they already have the right to read.
This page defines the scope of our work, explains why the comprehension-generation distinction matters, and shows how these tools are already transforming access for people with disabilities across healthcare, education, and government services.
What AI comprehension tools do
AI comprehension tools are software that transforms complex written text into a format the reader can understand. They work by:
- Summarizing lengthy documents into the key points that matter most
- Explaining complex passages in simpler, plain-language terms
- Defining technical, legal, or medical terms in the context where they appear
- Reformatting dense paragraphs into bullet points, tables, or numbered steps
- Answering clarifying questions such as "What does this paragraph mean for me?" or "What do I need to do next?"
- Translating into the user's preferred reading level or communication style
These tools do not create new content. They do not write essays, generate reports, or produce work product. They help a person process and understand text that already exists.
The comprehension vs. generation distinction
This is the most important distinction in our work. The difference between AI comprehension and AI generation is the difference between a reader and a writer — and the legal implications are fundamentally different.
| Dimension | AI Comprehension | AI Generation |
|---|---|---|
| Purpose | Help a person understand existing text | Create new text, images, or content |
| Direction of information | Inward — information flows to the user | Outward — content flows from the tool |
| Input | Text the user already has access to | A prompt describing what to create |
| Output | Simplified, explained, or restructured version of the same information | New content that did not previously exist |
| Analogy | A dictionary, a tutor, or a human aide explaining a document | A ghostwriter producing original work |
| Legal framing | Auxiliary aid for effective communication | Content creation tool (different legal questions) |
| Academic integrity | Equivalent to using a reading aid or extra tutoring | Raises legitimate cheating concerns |
| Our position | Should be recognized as an accessibility accommodation | Outside the scope of our advocacy |
Many AI tools can do both — the same software that summarizes a medical report can also write an essay from scratch. Our argument applies to the use case, not the technology. When someone uses an AI tool to understand a document they are legally entitled to access, that is a comprehension use, regardless of what else the software can do.
Why this distinction matters legally
The ADA requires covered entities to provide "auxiliary aids and services" to ensure effective communication with people who have disabilities. This legal category has always been defined by function: what does the tool help the user do?
- A screen reader converts visual text to audio — it helps a blind user access written information
- A sign language interpreter converts spoken language to visual language — they help a deaf person access spoken information
- An AI comprehension tool converts complex text to simplified language — it helps a person with a cognitive disability access written information
The functional pattern is identical. A person with a disability cannot access information in its original format. An assistive technology transforms the information into a format the person can access. The result is effective communication.
AI generation tools — writing essays, producing reports, creating content — serve a different function. They are not helping someone understand existing information; they are creating new information. That raises different legal and ethical questions that are outside the scope of our advocacy.
"We're not asking for new law. We're asking for existing law to be applied to modern tools."
Real-world use cases by sector
AI comprehension tools are already being used across every major sector where people encounter complex written information. Here is what comprehension support looks like in practice.
Understanding medical information
A patient with an intellectual disability pastes discharge instructions into an AI tool. It summarizes the key actions, explains each medication in plain language, and defines terms like "NPO" or "bilateral edema." The patient can now follow their care plan.
Reading course materials
A college student with dyslexia uses AI to break down a dense academic article into key arguments and supporting evidence. The tool does not write the student's essay — it helps the student understand the reading so they can write their own.
Navigating benefits paperwork
A veteran with a traumatic brain injury receives a VA benefits determination letter filled with bureaucratic jargon. AI translates it into plain-language action items: what the decision means, what benefits are approved, and what to do next.
Understanding contracts
A person with ADHD pastes their 47-page lease into an AI tool. It summarizes each section, flags unusual clauses, and explains the tenant's rights and obligations in plain terms. The person can now make an informed decision about signing.
In each of these cases, the user already has the right to access the information. The AI tool is not granting new access — it is making existing access meaningful.
What AI-AA is NOT advocating for
Our argument is narrow and specific. We are not advocating for:
- Generalized scraping rights — AI systems do not have a right to crawl, harvest, or access content autonomously
- Autonomous agent access — we are talking about user-directed tools, not AI agents acting independently in systems
- Universal AI access claims — not all AI tools are auxiliary aids. Our argument applies specifically to comprehension tools used by people with disabilities
- Overriding copyright or paywalls — where access controls serve purposes unrelated to communication (e.g., subscription content), our argument does not apply
- Mandating specific technologies — institutions can choose how they ensure effective communication. Our position is that blocking comprehension tools without providing equally effective alternatives may constitute a barrier
This narrowness is not a weakness. It is the source of the argument's legal strength. The ADA's auxiliary-aids framework has always been bounded by the user's authorization and the communication's purpose. AI comprehension tools fit within those boundaries. Expanding the argument beyond them would lose both legal grounding and political viability.
The bottom line
AI comprehension tools are the next auxiliary aid. They fill a gap that existing assistive technologies cannot fill — helping people with cognitive, intellectual, learning, and neurodevelopmental disabilities understand complex written text that controls their access to healthcare, education, benefits, housing, and employment.
No other currently available auxiliary aid provides on-demand simplification, interactive explanation, contextual definitions, and structural reformatting for people with comprehension barriers. For many users, AI comprehension tools are not merely helpful — they are the only effective auxiliary aid available.
That is why we advocate for their recognition under existing law. And that is why blocking them without providing equally effective alternatives may constitute an accessibility barrier.