| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 48.75 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Os LMS assumiram elevada importância nos últimos anos, num
contexto em que instituições de ensino de todo o mundo,
passaram a basear, parcial ou integralmente, o seu processo de
ensino. Somam-se a isso, os contextos recentes trazidos pelos
desafios da pandemia Covid -19 que levaram muitas escolas a
reestruturar os seus cursos para funcionarem com o suporte de
LMS. As interfaces dos sistemas de gestão de aprendizagem têm
sido avaliadas constantemente, a partir do uso de diferentes
métodos e abordagens, inclusive a partir do uso de avaliações
suportadas por técnicas de IA. Este trabalho procurou
compreender o estado do uso da IA no Design, especialmente, no
campo da avaliação de interfaces, procurando identificar qual é o
posicionamento que a IA está ocupando na prática dos
profissionais de design que avaliam e estudam interfaces
utilizadas para processo de ensino aprendizagem online.
Esta pesquisa discute e explora o papel da Inteligência Artificial
(IA) no Design no contexto atual. Para tanto, procurou identificar
os posicionamentos que estas tecnologias têm assumido nos
processos de design. A pesquisa situa-se no campo exploratório e
enquadra como objeto de estudo, a avaliação de interfaces
digitais de Sistema de Gestão de Aprendizagem denominados
como Learning Management System (LMS), avaliadas
tradicionalmente, a partir de heurísticas de usabilidade.
A pesquisa identificou que a IA pode contribuir para melhorar os
processos de criação e avaliação de interfaces, sem, no entanto,
afetar os princípios básicos do pensamento em Design, ainda que
a maneira como criamos, desenvolvemos e encontramos soluções
para os problemas esteja a se alterar com o uso desses recursos.
Por outro lado, foi possível reconhecer como os métodos de IA
ainda são limitados, com a sua capacidade de resolução de
problemas restrita aos contextos em que são aplicados, sendo
mais produtivo aplicá-los em situações em que a delimitação dos
problemas de design é bem definida.
Os desafios para adotar a IA no processo de avaliação de
interfaces são significativos, e ainda demandam um longa
trajetória a ser percorrida. Uma das necessidades constatadas por
essa pesquisa é aproximar os campos do Design e da IA, a partir
de ecossistemas que incentivem a colaboração entre designers e
cientistas de dados para criar soluções centradas no ser humano.
Diversos autores apontam para lacunas entre a formação
recebida nas universidades de Design e a prática dos designers,
sobretudo no que se refere ao conhecimento e uso dos recursos
da IA disponíveis para a execução de projetos. Apesar dos avanços
recentes no uso das tecnologias associadas à IA observado em
diversas áreas, a literatura evidencia que, na prática do design,
ainda há um conjunto de equívocos e desconhecimento sobre o assunto e sobre os possíveis impactos que a IA pode causar a
médio e longo prazos.
Um dos resultados da pesquisa foi o desenvolvimento de uma
metodologia a partir da sistematização de indicadores de
usabilidade com base nas heurísticas de Nielsen (Nielsen &
Molich, 1990) para contribuir para a simplificação do processo de
inspeção de interfaces. Além disso, uma sistematização de
indicadores de usabilidade para IA foi proposta e testes
experimentais com a adoção de modelos de IA foram realizados
para examinar a possibilidade de utilização de algumas tarefas de
IA para contribuir para o processo de inspeção de interfaces.
Quatro novas heurísticas foram propostas visando fazer face aos
novos desafios trazidos pelo contexto tecnológico vivenciado nas
últimas décadas.
A pesquisa permitiu evidenciar a necessidade de implementar
alterações na forma como os sistemas são desenvolvidos para
que a coleta de dados, alinhada ao respeito à privacidade e à
segurança dos utilizadores, seja considerada desde o princípio
para subsidiar a estruturação de conjuntos de dados que
poderão ser utilizados para alimentar modelos de IA,
adotados na inspeção da usabilidade das interfaces. Além
disso, observou-se a necessidade de promover o reforço de
boas práticas no desenvolvimento de sistemas, com vistas à
criação de projetos orientados à usabilidade. E por isso,
sugerimos e propomos novas formas de organizar os espaços
de criação de sistemas, orientadas à análise de usabilidade
com suporte da IA. E finalmente, apontamos a necessidade de
que trabalhos futuros sejam realizados para explorar ainda
mais os desafios para aproximar o Design e a Ciência de dados
para facilitar a prática da avaliação de interfaces.
This research discusses and explores the role of Artificial Intelligence (AI) in Design in the current context. To this end, it sought to identify the positions that these technologies have assumed in design processes, whether they are inserted either as design services or as material design. The research is located in the exploratory field and fits as an object of study, the evaluation of digital interfaces of Learning Management System (LMS), evaluated from usability heuristics. LMS have assumed great importance in recent years, in a context in which educational institutions around the world have started to base, partially or fully, their teaching process. Additionally, the recent contexts brought by the challenges of the Covid-19 pandemic, led many schools to restructure their courses to work with the support of LMS. The interfaces LMS have been constantly evaluated, based on the use of different methods and approaches, including the use of evaluations supported by AI techniques. The project sought to understand the state of use of AI in Design, especially in the field of interface evaluation, seeking to identify the position that AI is occupying in the practice of design professionals who evaluate and study interfaces used for online teaching and learning. The research identified that AI can contribute to improve the processes of creation and evaluation of interfaces, without, however, affecting the basic principles of thinking in Design, even though the way we create, develop and find solutions to problems is changing with the use of these resources. On the other hand, it was possible to recognize how AI methods are still limited and have their problem-solving capacity restricted to the contexts in which they are applied, making it more productive to apply them in situations where the delimitation of design problems is well defined. defined. The challenges to adopt AI in the process of evaluating interfaces are significant, and still require a long way to go. One of the needs identified by this research is to bring the fields of Design and AI closer together, based on ecosystems that encourage collaboration between designers and data scientists to create human-centered solutions. Several authors point to gaps between the training received from Design universities and the practice of designers, especially with regard to knowledge and use of AI resources available for the execution of projects. Despite the recent advances in the use of technologies associated with Artificial Intelligence observed in several areas, the literature shows that in the practice of design, there is still a set of misconceptions and lack of knowledge about the subject and about the possible impacts that AI can cause in the medium and long term. One of the results of the research was the development of a methodology based on the systematization of usability indicators based on Nielsen's heuristics to help simplify the interface inspection process. Furthermore, a systematization of usability indicators for AI was proposed and experimental tests with the adoption of AI models were carried out with the aim of examining the use of some AI tasks to contribute to the interface inspection process. Four new heuristics were proposed in order to face the new challenges brought by the technological context experienced in recent decades. This research highlights the need to implement changes in the way systems are developed so that data collection, in line with respect for users' privacy and security, is considered from the outset to support the structuring of data sets that can be used to feed AI models used in interface usability inspection. Furthermore, there was a need to promote the reinforcement of good practices in the development of systems, with a view to creating usability-oriented projects. And for that reason, we suggest and propose new ways of organizing the spaces for creating systems, oriented towards usability analysis with AI support. And finally, we point out the need for future work to be carried out to further explore the challenges of bringing Design and Data Science closer together to facilitate the practice of evaluating interfaces.
This research discusses and explores the role of Artificial Intelligence (AI) in Design in the current context. To this end, it sought to identify the positions that these technologies have assumed in design processes, whether they are inserted either as design services or as material design. The research is located in the exploratory field and fits as an object of study, the evaluation of digital interfaces of Learning Management System (LMS), evaluated from usability heuristics. LMS have assumed great importance in recent years, in a context in which educational institutions around the world have started to base, partially or fully, their teaching process. Additionally, the recent contexts brought by the challenges of the Covid-19 pandemic, led many schools to restructure their courses to work with the support of LMS. The interfaces LMS have been constantly evaluated, based on the use of different methods and approaches, including the use of evaluations supported by AI techniques. The project sought to understand the state of use of AI in Design, especially in the field of interface evaluation, seeking to identify the position that AI is occupying in the practice of design professionals who evaluate and study interfaces used for online teaching and learning. The research identified that AI can contribute to improve the processes of creation and evaluation of interfaces, without, however, affecting the basic principles of thinking in Design, even though the way we create, develop and find solutions to problems is changing with the use of these resources. On the other hand, it was possible to recognize how AI methods are still limited and have their problem-solving capacity restricted to the contexts in which they are applied, making it more productive to apply them in situations where the delimitation of design problems is well defined. defined. The challenges to adopt AI in the process of evaluating interfaces are significant, and still require a long way to go. One of the needs identified by this research is to bring the fields of Design and AI closer together, based on ecosystems that encourage collaboration between designers and data scientists to create human-centered solutions. Several authors point to gaps between the training received from Design universities and the practice of designers, especially with regard to knowledge and use of AI resources available for the execution of projects. Despite the recent advances in the use of technologies associated with Artificial Intelligence observed in several areas, the literature shows that in the practice of design, there is still a set of misconceptions and lack of knowledge about the subject and about the possible impacts that AI can cause in the medium and long term. One of the results of the research was the development of a methodology based on the systematization of usability indicators based on Nielsen's heuristics to help simplify the interface inspection process. Furthermore, a systematization of usability indicators for AI was proposed and experimental tests with the adoption of AI models were carried out with the aim of examining the use of some AI tasks to contribute to the interface inspection process. Four new heuristics were proposed in order to face the new challenges brought by the technological context experienced in recent decades. This research highlights the need to implement changes in the way systems are developed so that data collection, in line with respect for users' privacy and security, is considered from the outset to support the structuring of data sets that can be used to feed AI models used in interface usability inspection. Furthermore, there was a need to promote the reinforcement of good practices in the development of systems, with a view to creating usability-oriented projects. And for that reason, we suggest and propose new ways of organizing the spaces for creating systems, oriented towards usability analysis with AI support. And finally, we point out the need for future work to be carried out to further explore the challenges of bringing Design and Data Science closer together to facilitate the practice of evaluating interfaces.
Description
Keywords
Design Inteligência artificial Aprendizado de máquina UXD Educação