Publications

Permanent URI for this collectionhttps://repositorio.grial.eu/handle/123456789/34

Browse

Author: search.filters.author.Pereira, Juanan

Search Results

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    LAMB: Creando asistentes virtuales educativos sin programar
    (Servicio de Publicaciones Universidad de Zaragoza, 2025-06-11) Alier-Forment, Marc; Pereira, Juanan; Casañ, María José; García-Peñalvo, Francisco José
    LAMB (Learning Assistants Manager and Builder) es una plataforma de código abierto que permite al profesorado diseñar, desplegar e integrar asistentes de aprendizaje en sistemas como Moodle sin necesidad de programar. Estos asistentes se diferencian de los chatbots genéricos porque el docente controla su comportamiento, limita las respuestas a fuentes que él mismo selecciona y recibe aporta las citas correspondientes. LAMB se integra sin problemas en el LMS institucional via el estándar IMS LTI, mantiene los datos del alumnado dentro del marco de privacidad europeo y funciona tanto con modelos locales como con servicios en la nube. LAMB ofrece a los centros educativos una vía asequible y viable para incorporar tutorías basadas en IA sin perder la soberanía pedagógica ni la protección de los datos..
  • Thumbnail Image
    Item
    Safe AI in Education Manifesto. Version 0.4.0
    (2024-10-08) Alier-Forment, Marc; García-Peñalvo, Francisco José; Casañ, María José; Pereira, Juanan; Llorens-Largo, Faraón
    The Safe AI in Education Manifesto outlines ethical principles for integrating AI into educational environments. It emphasizes the need for human oversight, ensuring AI complements rather than replaces educators. Decision-making must remain transparent and appealable, protecting the educational process's integrity. Confidentiality is paramount; institutions must safeguard student data and ensure AI systems comply with stringent privacy standards. AI tools should align with educational strategies, supporting learning objectives without enabling unethical practices or adding complexity. The manifesto calls for AI systems to respect didactic practices, adapting seamlessly to instructional designs without burdening educators or students. It stresses accuracy and explainability, requiring AI outputs to be reliable, transparent, and verifiable. Interfaces must be intuitive, communicating their limitations to foster trust and critical engagement. Ethical training and transparency in AI model development are essential, including minimizing biases and disclosing data sources. The manifesto commits to advancing AI’s potential in education while prioritizing privacy, fairness, and educational integrity, providing a living framework adaptable to technological evolution. It can be signed at: https://manifesto.safeaieducation.org/