AI as Communication Wheelchair
There is a particular kind of person who thinks in cascades. The internal experience is rich, fast, and highly connected: one idea links to another, context piles on context, the relationships between things seem obvious from the inside, and the urgency of what needs to be said makes it hard to linearize the thought into the sequential form that language requires. When they speak or write without help, what comes out is often too compressed, too dense, or simply illegible to people who did not arrive at the thought the same way.
Other people hear this and say the person is a poor communicator. That is not quite right. The person is a poor translator. The experience they are having is precise. The translation apparatus is the problem.
Augmentative and Alternative Communication
The field of augmentative and alternative communication was built on a simple recognition: the inability to communicate through standard speech or writing does not mean the inability to communicate [1]. People with ALS, cerebral palsy, autism spectrum conditions, severe stuttering, or acquired brain injuries have thoughts, intentions, preferences, and needs that are fully real. The problem is the interface between the internal experience and the external form that others can receive. The solution is a different interface.
Light and McNaughton's foundational work in AAC showed that communication competence is not a single thing [1]. It involves linguistic competence, operational competence (how to use the tools), social competence, and strategic competence (how to manage the interaction). A person who lacks one of these does not lack all of them. A device or system that supports the deficient component enables the person's actual communicative intent to reach others.
The principle is straightforward but its application has been limited. AAC has primarily been developed for and associated with severe physical and neurological differences. The broader population of people who struggle not with gross communication capacity but with specific translation problems, those who cannot organize emotional language, those who think faster than they can speak, those whose natural communication style is illegible in mainstream social contexts, have been left largely without support.
The hypocrisy of the criticism
The most common criticism leveled at people who use AI to help them communicate is inauthenticity. The message was not really from them. They used a tool. The real person is hiding behind a machine.
This criticism has a specific and revealing inconsistency. The same people who make it would not say that a person's emails are inauthentic because they spell-checked them. They would not say that using a thesaurus makes an essay fake. They would not say that a speech therapist's coaching makes a presentation dishonest. All of these are tools that assist in the translation of internal intent into external form. The criticism of AI assistance is not a principled objection to communication support. It is a selective objection to this particular support, usually from people who find it threatening.
And the same people, in many cases, criticize the person's natural communication. Too blunt. Too emotional. Too dense. Too fragmented. They want the person to communicate differently, but they also want the person to do so without help. They demand a standard and penalize the tool that meets it.
We do not shame people for using wheelchairs. The same logic applies to using AI to translate intent into clearer speech.
Bauby and the primacy of intent
Jean-Dominique Bauby, the former editor of Elle magazine, suffered a massive stroke that left him with locked-in syndrome [2]. His entire motor system was paralyzed except for his left eyelid. Over months, he composed his memoir by blinking to select letters from an alphabetically arranged board. A transcriber recorded the selections. The resulting book, The Diving Bell and the Butterfly, is one of the most vivid and emotionally precise pieces of writing in recent memory.
No one questions whether that book was authentically Bauby's. It was. The tool that helped him externalize his intact interior experience did not make the experience less his. The tool made transmission possible. The transmission served the intent.
This is the correct frame for AI-assisted communication. When a person with unusual communication patterns uses AI to translate their intent into a form that other people can receive, the intent remains theirs. The translation is the service. The tool does not create the thought; it does not construct the relationship or the care or the meaning behind the communication. It builds a bridge between an interior that is real and an exterior that is legible.
Turkle's concern and its limits
Turkle has been a skeptical voice on this question, concerned that AI assistance in communication represents a kind of relational shortcut that substitutes performance for genuine engagement [3]. The concern is not trivial. There is a version of AI-assisted communication that is dishonest: using AI to manufacture affect that is not genuinely present, to perform care that is not felt, to create the impression of depth where there is none.
But this is a misuse of the tool, not a property of the tool itself. A wheelchair can theoretically be used in a performance designed to evoke sympathy that is not warranted. The tool's capacity for misuse does not determine its appropriate use. The appropriate use of a communication wheelchair is to support the transmission of genuine intent that is otherwise blocked by a translation problem.
The distinction matters. The person who uses AI to say things they do not mean is misusing the tool in a way that is dishonest. The person who uses AI to say more clearly and legibly what they genuinely mean and feel is not being dishonest. They are solving a translation problem.
What this asks of the people receiving the communication
It asks a modest recalibration. The question to ask of any communication is not "was this produced without tools" but "does this represent the person's genuine intent, and does it enable genuine relationship." If the answer to both is yes, the tool is doing its job and the criticism of its use is misplaced.
It also asks an acknowledgment of inconsistency. Most standard communication already involves extensive tooling: grammar checks, professional editors, communication coaches, PR handlers, legal review. These are accepted because they are familiar. AI assistance is resisted because it is new and because it democratizes access to support that was previously available only to those with money or institutional backing.
The person who writes an email with AI assistance and the person whose company employs a team of communication professionals to draft their messages are both using tools to improve their communication. The ethical question is not which tool but whether the intent is genuine.
The point
Language is not the thought. It is the vehicle for the thought. When the vehicle breaks down, the thought does not stop being real. Assistive technology for communication, in whatever form, exists to serve the transmission of genuine intent. AI is, for some people, the first tool that can actually do this. The appropriate response to a person using the tool that makes their genuine self accessible is not suspicion. It is the same acknowledgment you would give to anyone who found a way to finally be heard.
Sources
- Light, J., & McNaughton, D. (2014). Communicative competence for individuals who require augmentative and alternative communication. Augmentative and Alternative Communication, 30(1), 1-18.
- Bauby, J.-D. (1997). The Diving Bell and the Butterfly. Alfred A. Knopf.
- Turkle, S. (2011). Alone Together: Why We Expect More from Technology and Less from Each Other. Basic Books.