Los programas generativos “Transformer” AI, tipo ChatGPT ¿Una oportunidad para la evaluación formativa?

 


Imagen generada por DALL·E a partir de la frase "ChatGPT, digital art". 

Cada vez hay más evidencia de que las improvisaciones estadísticas de programas como  ChatGPT producen textos aparentemente creíbles pero que no tienen ninguna sustentación. El problema no es tanto que permitan el fraude en proyectos y artículos, como que hagan pasar por correctos lo que sólo son generación de textos "sin alma" [sin “beingness”] o simplemente escritos sin validar teórica o empíricamente, que se aceptan como ciertos acríticamente.

El aprendizaje no es dar cuenta sólo mediante exámenes y proyectos, de datos o de textos describiendo conceptos o hechos. Supone según los clásicos (Gagne, Merrill, Reigeluth) comprender, atribuir sentido, aplicar autónomamente lo aprendido y transferirlo a entornos distintos y cambiantes.

Para eso tiene que haber procesos de interacción y de retroalimentación. Y, si se utilizan proyectos en la evaluación, deben ser supervisados en su realización, en como se comprenden y en la atribución de sentido que el aprendiz hace.

Toda esta constelación de ideas, métodos y prácticas constituye, en el contexto del diseño instruccional, la evaluación formativa.

La evaluación formativa [formative assesment] es evaluación continua + feedback. En ese sentido debería haber actividades de evaluación continua que ofrezcan a los estudiantes comentarios formativos sobre su progreso y/o para ayudarlos a autoevaluar el desarrollo y la calidad del diseño

La evaluación para el aprendizaje (comúnmente conocida como evaluación formativa [formative assesment]) se centra en la retroalimentación tanto para profesores como para estudiantes.

En un mundo donde la IA aprende y nos supera en una lista creciente de dominios cognitivos, *Beingness* es el dominio final del esfuerzo intelectual humano y debe ser el núcleo de la educación. Y la evaluación formativa [formative assesment] debe ser un instrumento para ello.

Puede ver 

Los programas generativos “Transformer” AI, entre los que está ChatGPT, ¿una oportunidad para la evaluación formativa?


English version

The generative “Transformer” AI systems, an opportunity for formative assesment?  

There is an increasing evidence that the statistical improvisations of programs such as ChatGPT produce apparently credible texts but have no support. The problem is not so much whether they allow fraud in projects and articles, as that they some responses are assumed as correct when are only the generation of "soulless" texts [without "beingness"] or just written without theoretical or empirical validation, which are accepted as uncritically true.

Learning is not giving an account only through exams and projects, data or texts describing concepts or facts. According to the classics (Gagne, Merrill, Reigeluth), it requires features as understanding, attributing meaning, autonomously applying what has been learned and transferring it to different and changing environments.

For that reason, there must be a processes of interaction and feedback. In addition, if projects are in progress, they must, in the evaluation, be supervised in their realization, in the understanding and in the attribution of meaning that the apprentice makes.

All this constellation of ideas, methods and practices constitutes, in the context of instructional design, formative assessment.

Formative assessment is continuous assessment + feedback. In this sense, there should be continuous evaluation activities that offer students formative feedback on their progress and/or to help them self-assess the development and quality of the design.

Assessment for learning (commonly known as formative assessment) focuses on feedback for both teachers and students.

In a world where AI learns and outperforms us in a growing list of cognitive domains, *Beingness* is the ultimate domain of human intellectual endeavor and should be at the core of education. Moreover, formative assessment [formative assessment] must be an instrument to achieve this goal.

 

  

Miguel Zapata Ros

 

He is Honorary Professor of the Department of Electromagnetism and Electronics at the University of Murcia. Additionally, he is a member of the Information and Knowledge Engineering PhD program at the University of Alcalá and a researcher at the Interuniversity Institute of International Economics of the Universities of Alicante, Valencia and Jaime I of Castellón.

Dr. Zapata got his PhD in Computer Science from the University of Alcalá, after getting his Master's Degree in Educational Theory from the University of Murcia and a BA in Mathematics from the University of Valencia.

His main job has been as a professor of Education, specialized in Mathematics.

Currently, he is an international evaluator of the A3ES-Agência de Valiação e Acreditação do Ensino Superior de Portugal.

His honored recognition has led him to become a Representative at the New York headquarters and Member of the International Council (Geneva headquarters) of the INTERNATIONAL COMMISSION ON DISTANCE EDUCATION which, since 2003, has consultative status with the UNITED NATIONS Economic and Social Council.

Besides all these extraordinary achievements, Dr. Zapata is the Editor and founder of RED Revista de Educación a Distancia (WOS, Scopus indexed) since 2001.

He has also authorized more than 200 publications on teaching, learning, ICT and on teaching and university policy in the Knowledge Society in refereed and indexed publications.

He has been a visiting professor and advisor to universities in the US, UK, Uruguay, Argentina, Mexico, Chile, Colombia, Peru, Ecuador, Venezuela, Portugal, Turkey and Cyprus. Advisor to ANEP Uruguay and the Telefónica Foundation of Spain, Uruguay and Argentina, among others.

And Advisor and technical adviser to the scientific journals of University Teaching and Law Education

All these merits in recognitions have gained him the recognition not only in the fields of eduation and technology, but he was also recognized by the Association of Ecuadorian University Faculties of Medicine which granted him its Recognition of Institutional Merit AFEME 2017

 

 --------------

(1) Resumen de mi intervención en 13th WORLD CONFERENCE ON EDUCATIONAL TECHNOLOGY RESEARCHES (WCETR-2023). University of Barcelona. Barcelona, Spain. March 29-30, 2023

 



Comentarios

  1. Desde que comencé a estudiar acá https://cursos-gratis.com.es/c-curso-gratis-e-commerce-y-social-media debí establecer una rutina diaria definir un horario fijo para estudiar, descansar y realizar otras actividades, de modo que la jornada tenga una estructura similar a la de una clase presencial.

    ResponderEliminar

Publicar un comentario

Entradas populares de este blog

Donde se concluye que la depredación es una consecuencia de las políticas de Acceso Abierto sin que se haya conseguido nada a cambio, al menos algo de las expectativas iniciales*.

Edición híbrida (I): Crossciting*.

Recursos educativos, contenidos y libros de texto con la INTELIGENCIA ARTIFICIAL GENERATIVA