International Journal of Latest Research in Humanities and Social Science (IJLRHSS) Volume 04 - Issue 01, 2021 www.ijlrhss.com || PP. 26-33 26 | Page www.ijlrhss.com Diagnostic for the design of a virtual educational assistant for visually impaired children Erick Ibarra Cruz 1 , Dulce María Flores Olvera 2 1 Benemérita Universidad Autónoma de Puebla, Facultad de Filosofía y Letras, Reforma 410, Puebla, México 2 Benemérita Universidad Autónoma de Puebla, Facultad de Filosofía y Letras, Reforma 410, Puebla, México Abstract: This paper describes a guide for developing a prototype of an intelligent virtual assistant (IVA) to aid in the teaching-learning process of children that are visually disadvantaged. The recommendations put forth by experts that participated in the study give direction concerning the characteristics that should be considered in the design of the software agent so that it can interact with a visually disadvantaged user, taking into account her sensorial deficiencies and special education needs through a personalized IVA that functions as a virtual tutor. Currently, in the context of research and application of this technology, there is no software environment adapted to the needs of these students. With this study, we explore how artificial intelligence can be integrated into learning environments with blind and visually deficient elementary school students. This paper presents a general vision of ongoing work aimed at constructing a prototype of the IVA system. Keywords: intelligent virtual assistant, artificial intelligence, visual deficiency, teaching-learning. Introduction Currently, it is increasingly common to hear about Intelligent Virtual Assistants (IVA) present in last generation cell phones, laptops, tablets and desk computers, and intelligent speakers or watches. For example, to cite some of the most well-known (Lamontagne, Laviolette, Khoury, Bergeron-Guyard, 2014; Ibarra, 2020), Apple equipment comes with factory installed personal assistance software called Siri. Smart phones with Android operative systems are installed with Google Assistant developed by Google. In recent technology devices with the operative system Microsoft Windows comes the assistant Cortana, while Samsung uses the assistant S Voice or Bixby, and Amazon speakers use the assistant Alexa. A virtual assistant is a software agent or program that can interpret human speech and respond with synthesized voices using artificial intelligence techniques, such as natural language recognition, and can interact with the system through voice commands. For this reason, these assistants are also known as voice assistants. These programs can work with another program or with a user and assist through a conversational interface to perform tasks, from asking simple questions about a particular topic, reproducing music or multimedia, administrating email, giving reminders of pending tasks, to manipulating home automized devices, for example turning on lights, activating and deactivating security cameras, doors and window, turning air conditioning and heating on and off, and communication, it can pose a security and privacy risk (Hoy, 2018; Chung & Lee, 2018; Hasan, Shams, Rahman, 2020). The interaction that is produced between a virtual assistant and a person must be through use of key words or short phrases expressed orally in natural language. The virtual assistant then in listening mode captures the message, processes it, interprets the users request, and gives an answer in a synthetic voice that varies in function of the programming and system configuration. One of the most important characteristics of the IVA is their great educational potential in the field of visual impairment because it offers the student the possibility of communicating with the device without the need to write with the keyboard or use Braille, so widely used by this community (De Freitas, Martins, Rabello, Rodrígues, Monteiro, 2009). Nor is it necessary to use any other technical aid for the blind; it only requires voice commands. Thanks to the artificial intelligence (AI) behind the design and programming of the system, natural language recognition makes it possible to execute functions that the use requests, such as search for information in the Internet, call a contact by telephone, perform simple mathematical operations, and even tell a joke (Hoy, 2018). The multinational company IBM, pioneer in the development of virtual assistants, created the virtual assistant Watson, a system that allows creating, training and implementing conversational interactions in any application, device or channel. An important characteristic is that it uses two methods of automatic learning of artificial intelligence called Machine Learning and Deep Learning. This program has been tested very successfully in academic programs in university teaching and functions as an assistant, helping students in private tutoring. Moreover, the assistant follows up on student progress, identifying where there are gaps in