HumanAI at INTERACCIÓN 2026: Cognitive Accessibility in Generative AI

The HumanAI project participated in the XXVI International Conference on Human-Computer Interaction (INTERACCIÓN 2026), held in Albacete, Spain, with contributions from the two teams of the coordinated project: HULAT at Universidad Carlos III de Madrid (UC3M) and NIL at Universidad Complutense de Madrid (UCM).

Two research papers were presented. “Oversight Mechanisms for Cognitive Accessibility in LLM-Based Generative Interaction,” by Lourdes Moreno, Paloma Martínez and Blessing Jerry, proposes mechanisms for integrating human oversight into generative systems through observable signals, rules and triggers, Human-in-the-Loop and Human-on-the-Loop supervision, and monitoring through KPIs. The work includes examples based on multi-agent workflows and the verification of generated accessible interfaces.
The paper “Bridging Co-Design and Prototyping: Impact–Effort Prioritization of Cognitively Accessible Generative AI Designs,” by Raquel Hervás and Virginia Francisco, presents a method for selecting design proposals obtained through co-design with people with intellectual disabilities, combining their expected impact with the technical effort required for prototyping.

HumanAI was also presented in the Research Projects session through “Project HumanAI: Towards Cognitively Accessible Generative AI Tools and Generative-AI-Driven UI and Content,” by Paloma Martínez, Lourdes Moreno, Raquel Hervás and Virginia Francisco. The presentation introduced the project’s two complementary research directions: making generative AI tools cognitively accessible and using generative AI to produce accessible content and user interfaces.