Reflections from ATIA 2026 – AI and the Future of AAC Symbol-Based Communication

Returning, once again, from the annual Assistive Technology Industry Association Conference in Orlando, one takeaway stood out more than any single product launch or keynote address and it wasn’t the subzero temperatures. Artificial intelligence was everywhere, but not in the way headlines might have you believe.

Rather than dominating the agenda as a standalone theme, AI was seamlessly woven into the fabric of assistive technology: embedded in AAC platforms, captioning tools, voice technology, visual aids, configuration processes, and analytics dashboards. It emerged in technical presentations, on the expo floor, and most tellingly, in informal conversations among professionals, AAC users, developers, families, and researchers. 

For those of us at Global Symbols, this shift is significant. It signals a move beyond novelty toward thoughtful integration—and prompts us to consider the impact of automation on communication, identity, and agency.

A conference defined by questions, not hype

ATIA is always busy, but the atmosphere this year felt particularly reflective. Walking through the Marriott Hotel, the most common phrases overheard were not about algorithms or models, but about outcomes:

• “It saves teachers time.”
• “It adapts to how the user communicates.”
• “Setup is much quicker now.”
• “It works better in noisy rooms.”

The language of AI was present, but often implicit. What people were really discussing was reduced burden, faster customisation, and smoother daily use. In assistive technology, that framing is revealing. The community does not adopt tools because they are technically impressive; it adopts them because they are dependable, understandable, and respectful of the people who rely on them.

In AAC-focused discussions and presentations often the tone was pragmatic rather than speculative. Speakers discussed integration with existing platforms, evaluation in real-world settings, and the complexities of deployment in schools, communities and clinics.

Technical advances were quickly followed by practical questions:

• How transparent is this for users and families?
• Can users override it easily?
• What happens when predictions are wrong?
• How is personal data handled?

In some cases sophisticated features were not immediately implemented, because users did not yet trust it, even though it performed well in testing. That moment captured a recurring theme across the conference: progress in assistive technology is shaped as much by relationships and confidence as by technical capability.

Implications for symbols and AAC

From a symbol-based AAC perspective, many of the emerging applications of AI are highly relevant:

• automated vocabulary suggestions
• faster creation of communication boards
• multilingual symbol workflows
• personalisation based on use patterns
• tools to support professionals’ analysis and reporting

These possibilities are exciting, but they also sharpen long-standing questions in AAC design.

A recurring conversation concerned authorship. When a system predicts a phrase or proposes a sentence structure, whose voice is being expressed? The individual’s? The AI model’s? The designer’s assumptions or perceptions as embedded in training data?

For people who communicate using symbols, that question is fundamental. AAC is not merely about efficiency; it is about self-expression, identity, and control. Automation that accelerates communication must still leave the user firmly in charge.

In workshops attendees spoke about the promise of AI to reduce administrative workload and speed up configuration, alongside worries about over-standardisation. If systems are trained primarily on “typical” language patterns or dominant cultural contexts, they risk narrowing rather than widening the range of expression available to AAC users.

That tension, between efficiency and individuality, is precisely where symbol systems must remain anchored.

Seeing the wider ecosystem

One of ATIA’s strengths is that it brings together the full spectrum of assistive technology. In a single walk through the hall, it was possible to encounter:

• speech-to-text and captioning platforms
• synthetic and personalised voice technologies
• computer-vision systems for reading and navigation
• environmental access tools
• dashboards designed for therapists and educators

This breadth highlights how symbol-based AAC increasingly sits within a larger technological landscape. Multiple automated systems may now operate together: prediction layered onto speech output, transcription feeding into reports, analytics shaping intervention plans.

That convergence can be powerful, but it also increases complexity. Several of the most thoughtful discussions focused not on headline features, but on what systems expose to practitioners: logs, explanations, configuration options, and ways to understand why a tool behaved as it did.

Inclusion, in this context, becomes a technical discipline. Transparency, inspection, and control are not optional extras; they are prerequisites for adoption.

The conversations that mattered most

Some of the most revealing exchanges did not take place in lecture theatres or formal demos. They happened in coffee queues, beside charging stations, or sitting on the benches and around tables between sessions.

• A parent worried about where recordings of their child’s voice were stored.
• A school technology lead sharing the challenge of explaining automated features to families already navigating complex systems.
• A researcher considering the risk of cultural flattening if symbol-generation tools draw from narrow datasets.

None of these voices were dismissing AI. They were asking for it to be developed with care, accountability, and humility. That insistence on responsibility, shared across disciplines and roles, felt like one of the strongest signals to come out of this year’s conference.

What felt different this year

Looking back, three shifts stood out.

AI is becoming infrastructure.

Rather than sitting in a separate “future tech” category, automation now appears embedded in everyday tools and workflows. That is a marker of maturity.

Ethics has become operational.

Debates were framed in classroom-level terms: does it work offline, can it be disabled easily, how do you explain it quickly, does it perform for atypical speech?

Trust is the central currency.

The most compelling exhibitors were not those promising dramatic transformation, but those willing to discuss limits, controls, and fallback options. In assistive technology, honesty builds confidence.

Leaving Orlando with renewed focus

As the final afternoon wound down and crates were sealed for shipping home, a few phrases kept resurfacing:

As the final afternoon wound down and crates were sealed for shipping home, a few phrases kept resurfacing:

• AI as scaffolding.
• Prediction versus authorship.
• Designing for outliers, not averages.

That last point feels especially relevant for Global Symbols’ work. People who rely on AAC often sit at the edges of datasets and interface assumptions. If AI systems are trained only on dominant languages or cultural norms, they risk reinforcing the barriers assistive technology exists to dismantle.

Global Symbols’ long-standing commitment to openness, multilingualism, and diverse representation aligns closely with the direction many at ATIA are calling for. As automation becomes more embedded in AAC systems, the need for transparent, culturally responsive, and user-controlled symbol resources only grows stronger.

ATIA this year suggested that the field is ready for that challenge.

AI is not something happening to assistive technology. It is something being shaped within it, by all stakeholders, including users, working together.

The task ahead is to ensure that these tools become not just more powerful, but more accountable, more inclusive, and more responsive to the people whose voices depend on them.

That is where symbol-based communication has a crucial role to play.