Rethinking language access in AAC
Artificial Intelligence (AI) is becoming part of everyday communication. For AAC users, AI-powered tools can offer meaningful benefits, from realistic synthetic voices to supporting with spelling, abbreviation expansion and grammar correction. As we introduce more of these tools into Grid, we are keen to engage with users, clinicians and the wider AAC community to get a well-rounded view of how this technology can be used in practice. In this blog, Speech and Language Pathologist Brenda Del Monte explores how emerging technologies like AI are opening up new possibilities for language access in AAC, and prompting us to think about how communication is supported. For decades, the field of AAC has worked tirelessly to give individuals access to language. Devices have evolved. Systems have improved. Research has guided us. And yet, beneath all of this progress, there is a quiet tension that continues to shape practice in ways we may not always recognise. It is the tension between being a gatekeeper and becoming a gate opener. At its best, AAC is about access. Access to communication. Access to connection. Access to identity, autonomy, and belonging. But somewhere along the way, in our effort to make communication measurable, teachable, and systematic, we unintentionally began to control it. We built systems filled with words. Carefully selected words. Researched words. High-frequency words. Words that we, as professionals, determined would be most useful. This is not inherently wrong. In fact, it was built with the best of intentions. Research has shown that a relatively small set of words, often referred to as core vocabulary, makes up a large percentage of everyday language use (Banajee, Dicarlo, & Stricklin, 2003). These words, like “go,” “more,” “like,” “it,” “that,” are flexible and generative. They can be used across contexts. They form the backbone of language. So we set out to teach them. We tracked them. We measured them. We created goals around them. We built therapy plans designed to increase their use with accuracy and consistency. In response to these limitations, many professionals have intentionally moved from limited systems toward robust AAC systems, placing 60, 84, or more words on a single page and modeling language consistently throughout the day. This shift reflects an important understanding that language should be available, not withheld. However, even with this expanded access, something is still missing. We are not consistently seeing the language growth we expected. Why? Because access alone, without autonomy, is not enough. Even when large amounts of language are available, the way we use those systems often remains rooted in compliance. We ask students to find specific words. We prompt predetermined responses. We take data on accuracy. We measure whether the “correct” word was selected. But how can we determine accuracy when we cannot access the intent behind the message? Even within a robust system, language can still be restricted. Even within a system designed to expand communication, we may still be gatekeeping without realising it. To understand why, we need to look at how language actually develops. When we look at how language develops in children who speak, we see something very different. Language is not taught through isolated word mastery or restricted exposure. It is developed through interaction, shared experiences, and exposure to language that often exceeds the child’s current level of use (Vygotsky, 1978; Nelson, 2007). We do not restrict children to a predetermined set of words before allowing access to others. We do not require mastery before expansion. We do not design their language experience around what is easiest to measure. Instead, we follow their interests. We respond to what captures their attention. We build language around what excites them, what comforts them, what makes them curious. And those words are the ones tied to interest and experience which are often what we classify in AAC as fringe vocabulary. Words like vacuum, garage door opener, train, YouTube, grandma’s house. Words that may not appear frequently across all users, but appear constantly for that individual. Words that carry meaning. Words that invite interaction. Words that spark curiosity and language. And yet, when we limit access to those words in favour of what is most measurable, we begin to shift from supporting communication to shaping it from the outside. This concern aligns with longstanding cautions in AAC research regarding over-reliance on structured prompting and facilitator influence (Beukelman & Mirenda, 2013; Kent-Walsh & Binger, 2010). This is where the role of gatekeeper emerges. Gatekeeping in AAC rarely looks intentional. It often sounds reasonable. It sounds like, “There are too many buttons.” It sounds like, “This is too overwhelming.” It sounds like, “We need to start simple.” But the impact is significant. Because every time we reduce vocabulary, delay access, or restrict language exposure, we are making a decision about what words someone is allowed to have. We are deciding what is worth saying before they have the chance to say it. We are filtering language through our own experiences, our own assumptions, and our own need to measure progress. And in doing so, we risk limiting not just communication, but identity. A gate opener does not reduce language. They expand it. A gate opener recognises that interest drives motivation, and motivation drives language development. For the child who is fascinated by vacuum cleaners, vacuum is not fringe. It is central. It is the entry point into questions, comments, storytelling, humour, and connection. “What is that part?” From a single “fringe” word, an entire language system can grow. When we follow interest, we follow meaning. And meaning is what makes language stick. This is also where the concept of co-authorship becomes important. This is where emerging tools, including AI, offer a powerful shift. Historically, AAC systems have been limited by the vocabulary that could be manually programmed. Even the most robust systems are still reflections of what someone else thought to include. They are shaped by human experience, human bias, and human limitation. Co-authorship is not new. In fact, all communication is, in some way, co-authored. Our ideas are shaped through interactions, relationships, conversations, and exposure to the world around us. This article itself is co-authored, not only by the writer, but by years of conversations with colleagues, families, researchers, and individuals who use AAC, as well as the collective knowledge of the field. The goal, then, is not to eliminate support. It is to ensure that support does not override authorship. In AAC, co-authorship becomes problematic when the partner or system determines the message. But it becomes powerful when the communicator remains the source of intent, using available tools to expand expression. This is where emerging tools, including AI, offer a meaningful shift. AI does not need to replace the communicator to be effective. When used intentionally, it can expand access to language in ways that were previously limited by what could be manually programmed. It can introduce new vocabulary, suggest phrasing, and reflect patterns of language that extend beyond a single user’s environment. Importantly, the AAC user remains in control. They select, reject, modify, and shape the output. This is co-authorship in its most ethical form. Not replacement. Not prediction of thought. But expansion of possibility. Research on communicative competence reminds us that true communication involves more than operational skill. It includes linguistic, social, and strategic competence, all working together to support meaningful participation (Light & McNaughton, 2014). When we become gate openers, we support all of these areas. We begin to ask different questions. Instead of “What words should we teach next?” we ask, “What is this person trying to say?” Instead of “How do we measure accuracy?” we ask, “How do we expand expression?” Instead of “Is this too much?” we ask, “What would it look like to trust potential?” We move towards expanding possibility. And as these questions shift, it can feel like everything we have built in AAC is being questioned. If we are not relying on core word targets in the same way, if we are not measuring accuracy in the same way, if we are expanding vocabulary instead of limiting it, then what are we using to guide intervention? The answer is not nothing. The answer is that we already have strong, well established frameworks. We simply need to apply them in ways that align with access, authorship, and authentic communication. The SETT framework developed by Joy Zabala reminds us to consider the Student, Environment, Tasks, and Tools when making decisions (Zabala, 2005). It was never designed to reduce options. It was designed to ensure that tools match the individual. When we apply SETT through the lens of gate opening, the question shifts from “How do we simplify the system?” to “What supports allow this individual to express what they want to say in this environment?” The framework stays the same. The interpretation evolves. We can also continue to ground our work in communicative competence. Communication extends beyond word production to include linguistic, operational, social, and strategic competence. Expanding access to language directly supports these areas by increasing opportunities for expression, participation, and repair (Light, 1989; Light & McNaughton, 2014). In the same way, AAC intervention can be examined through the expansion of communicative functions. Instead of measuring whether a specific word was used correctly, we can examine whether the individual is requesting, commenting, asking questions, protesting, sharing ideas, or engaging socially. This is data. It is just different data. It is data that captures participation, intent, and growth in communicative competence rather than accuracy on isolated words. It is data that reflects communication, not compliance (Burkhart & Porter, 2019). We are not abandoning structure. We are refining it. And this refinement matters. Research in AAC and literacy further emphasises that language and literacy development are driven by access to rich, meaningful language experiences, not restriction to limited vocabulary sets. When opportunities are reduced, outcomes are often reduced, not because of learner capacity, but because of limited access to language and instruction (Erickson & Koppenhaver, 2020). This shift from gatekeeper to gate opener also requires us to confront an uncomfortable truth. Many of the practices we rely on feel safe because they are measurable. Tracking accuracy is measurable. Limiting vocabulary is manageable. But communication itself is not easily reduced to numbers. It is dynamic, relational, emotional, and deeply human. When we prioritise what is measurable over what is meaningful, we risk missing the very thing we set out to support. No one enters this field to limit language. We enter it to unlock it. Being a gate opener aligns with that belief and purpose. It means presuming competence not only in what is visible, but also in what has not yet been expressed. It means presuming potential, recognising that language can grow in ways we cannot fully predict. It means presuming intellect, trusting that the person in front of us has something to say, even if the system has not yet made it easy for them to say it. And creating environments where language is not restricted by our expectations but expanded by possibility. When we open the gate, we do more than increase vocabulary. We increase autonomy. We strengthen identity. We support participation. We invite individuals into conversation not as responders to limited choices, but as authors of their own thoughts. Maybe this is the moment where we pause and remember why we chose this work in the first place. Not to decide what someone should say before they’ve had the chance to say it. But to make sure every person has access to their voice. The curiosity is already there. The desire to learn new things is already there. The thoughts are already there. The preferences are already there. The personality, the humour, the opinions, the stories are already there. Not perfect sentences. Not accurate data points. Not mastery of a predefined list. But authentic expression. Our job is not to build the voice. Our job is to remove the barriers that are in the way of it. When we shift from gatekeeper to gate opener, we stop asking, “What language should we allow?” and start asking, “What might be possible if we didn’t limit it?” Because the goal was never compliance. The role of the clinician, then, is not to decide what language is allowed, but to ensure that language is accessible. To build systems that are rich, flexible, and responsive. To use tools, including AI, in ways that expand rather than replace. To follow interest, not suppress it. To trust that when given access to meaningful language, individuals will use it. And in doing so, we shift the field. From gatekeeping to gate opening. From limitation to possibility. From programmed words to authentic voice. Because in the end, communication is not something we give. It is something we make possible. It was always about connection. And when we open the gate wide enough, we don’t just see more words. We see the person. Banajee, M., Dicarlo, C., & Stricklin, S. B. (2003). Core vocabulary determination for toddlers. Augmentative and Alternative Communication, 19(2), 67–73. https://doi.org/10.1080/0743461031000112034 Beukelman, D. R., & Mirenda, P. (2013). Augmentative and alternative communication: Supporting children and adults with complex communication needs (4th ed.). Paul H. Brookes Publishing Co. Burkhart, L. J. (2004). Assistive technology and AAC: Principles for supporting communication. (You can also list as unpublished training materials or professional resources if needed depending on publication requirements) Erickson, K. A., & Koppenhaver, D. A. (2020). Comprehensive literacy for all: Teaching students with significant disabilities to read and write. Paul H. Brookes Publishing Co. Kent-Walsh, J., & Binger, C. (2010). Teaching partners to support the communication of individuals who use AAC: A systematic review. American Journal of Speech-Language Pathology, 19(2), 97–116. https://doi.org/10.1044/1058-0360(2009/09-0034) Light, J. (1989). Toward a definition of communicative competence for individuals using augmentative and alternative communication systems. Augmentative and Alternative Communication, 5(2), 137–144. https://doi.org/10.1080/07434618912331275126 Light, J., & McNaughton, D. (2014). Communicative competence for individuals who require augmentative and alternative communication: A new definition for a new era of communication? Augmentative and Alternative Communication, 30(1), 1–18. https://doi.org/10.3109/07434618.2014.885080 Nelson, K. (2007). Young minds in social worlds: Experience, meaning, and memory. Harvard University Press. Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press. Zabala, J. S. (2005). Using the SETT framework to level the learning field for students with disabilities. Zabala & Associates.
How language is shaped in AAC systems
How language develops
Contrast this with the role of a gate opener
“It’s loud.”
“Turn it on.”
“I like that one.”
“My house has one.”How AI tools can help us become gate openers
Applying established frameworks through a new lens
Opening the gate to authentic voice
References
More about AI tools in Grid:
