Artificial Intelligence is often celebrated for its ability to optimize industrial processes or generate creative content. But what happens when we place this extraordinary power at the service of vulnerability and inclusion?
For a person with intellectual disabilities and their entire support network (family, school, clinicians, and associations), AI ceases to be a mere tech trend and becomes a tool capable of enabling a fundamental right: self-determination. However, general-purpose models carry biases and hallucination risks that, in critical contexts, turn into hazards rather than opportunities.
In this talk, drawing from my dual perspective as an engineer and father of a daughter with Down Syndrome, we will explore how to transform AI from a simple oracle into a specialized assistant for independent living, analyzing the technology’s potential across four key domains:
Language and Communication Accessibility Machine Learning and Adaptive Interfaces: how personalized models trained on specialized datasets can translate non-standard speech, vocalizations, or behavioral patterns into concrete actions, giving voice to those who struggle to communicate with the outside world.
Building Knowledge Bases for People with Intellectual Disabilities, Caregivers, and Professionals It is essential to build a “safe memory” for the support network. We will see how AI can securely process clinical, pedagogical, and regulatory documentation to provide verified answers, supporting therapists and families. Furthermore, research findings can be made accessible, providing visibility into scientific advancements across diverse fields.
Autonomy in Daily Life Daily living autonomy assistants: developing autonomous agents capable of simplifying daily complexity, giving individuals with intellectual disabilities that extra bit of help that enables a more independent life while easing caregiver burden.
Democratizing Information Cognitive Simplification: using Large Language Models for automated translation of complex texts (laws, contracts, manuals) into Easy-to-Read formats (easy to read and understand), guaranteeing the right to information and eliminating cultural isolation.
The Crossroads: Radical Inclusion or a New Digital Divide? Addressing this topic today is not optional. Ignoring cognitive accessibility in the development of future Artificial Intelligence systems risks creating an even wider social and economic divide, permanently excluding the most vulnerable people from an increasingly automated world. Conversely, balancing AI Act compliance with the need for specific, inclusive datasets is the only path toward true technological democratization. The challenge before us is not merely technical, but profoundly cultural and ethical. AI can unlock extraordinary doors, but we must choose the direction: my invitation is to focus our energy, sensitivity, and human intelligence toward a single common goal, which is the only one that truly matters: building a better, fairer world that leaves no one behind.
