AI Transforming Assistive and Inclusive Technologies

Equitable AI Alliance

Introduction

Artificial Intelligence (AI) is reshaping the landscape of assistive and inclusive consumer technologies, creating a profound shift in how people with disabilities access, interact with, and benefit from digital tools. What once required specialised devices explicitly designed to compensate for functional limitations is increasingly delivered as assistive features and functions within mainstream technologies that adapt dynamically to individual needs. This evolution marks a transition from fixed, prescriptive tools to flexible systems that respond to context, preference, and behaviour.

From Specialised Tools to Adaptive Consumer and Educational Technologies

Historically, assistive technology consisted of highly targeted solutions such as screen readers, communication devices, and alternative input systems. These tools enabled access but often required users to adapt to and learn to use the technology rather than the technology adapting to them. AI has altered this balance. Modern systems can interpret speech patterns, learn vocabulary, adjust to motor behaviours, and present information in personalised ways.

As AI capabilities expand, mainstream technologies have absorbed functions once considered the exclusive domain of assistive tools. Smartphones, tablets, smart speakers, and wearable devices now include sophisticated accessibility features such as speech recognition, captioning, magnification, environmental description, adaptive user interfaces, and real-time translation.

Emerging Risks and Challenges

Despite the promise and delivery of AI-enabled accessibility, important risks must be confronted. Algorithms trained on limited datasets may misinterpret or fail to understand the speech, expressions, movements, or behaviours of disabled learners. Individuals who do not conform to notions of a ‘typical’ user are at increased risk of misrecognition, exclusion, or inaccurate assessment.

In both educational and workplace settings, AI-driven monitoring tools can introduce further challenges. Systems that track engagement, attention, behaviour, or sentiment may wrongfully penalise learners with disabilities whose cues differ from assumed norms, raising concerns regarding autonomy, dignity, and fairness.

Accessibility barriers also persist within many AI-enabled interfaces. While the functions may promise inclusion, the user experience can still present obstacles for screen reader users, individuals with cognitive access needs, or those requiring customised interaction. Rapid software updates compound these difficulties, disrupting established workflows or removing essential features.

Policy Evolution for Inclusive AI

To fully realise the opportunities of AI while mitigating its risks, policy and provision frameworks must evolve. A modern understanding of assistive and accessible technology must recognise that accessibility is no longer confined to specialised devices. Instead, mainstream consumer products increasingly form part of the assistive ecosystem, requiring new approaches to funding, procurement, and assessment.

Equally, the successful implementation of AI-enabled technologies depends on the capabilities of the workforce supporting them. Educators, therapists, technologists, and support workers require practical training not only in the operation of AI tools but also in ethical and responsible use, accessibility standards, and methods for matching technology to individual needs. Moreover, the ready availability of AI assistive features and functions within mainstream technologies facilitates much greater self-determination of solutions by learners with additional needs. A rigid dependency on professionals is challenged as learners make their own choices about how their needs are best met.

These changes introduce new expectations regarding equity. Access to devices, connectivity, and digital literacy must be treated as essential components of an inclusive education and social system. Without robust public policy addressing affordability, discriminatory pricing, and digital exclusion, the benefits of AI risk being unevenly distributed.

Sustainability is a critical dimension. Effective implementation of AI-enhanced technologies requires careful lifecycle planning, stable update pathways, backward compatibility, and clear mechanisms for maintenance, repair, and replacement.

Assistive technology services must also rethink their operating models. Assessment should increasingly focus on context, environment, and individual preference rather than impairment categories or device types. Continuous assessment better reflects the dynamic nature of AI tools.

Designing for a Future Guided by Human Values

Looking ahead, a sustainable and inclusive future for AI in assistive and consumer technologies depends on recognition that human values must guide technological change. Independence, dignity, autonomy, transparency, and co-creation should shape the way AI is developed and deployed.

AI will not, by itself, determine the future of assistive and inclusive technologies. The outcomes will depend on the collective choices of governments, developers, educators, service providers, and disabled people themselves.

Written for OEB 2025 by David Banes.

Meet David on Thursday for his presentation: “Building and Implementing Inclusive AI for All – Addressing the Needs of Learners With a Disability“