Work

Towards Improving Personal Assistants and Educational Software: How Questions Are Answered Affects Learning .

Year: 2021
Type: article
Abstract: Using technologies like artificial intelligence (AI), machine learning, natural language processing and voice recognition, personal assistants and other query-answering systems have made enormous progress in understanding people's questions, retrieving relevant information and using that information to answer people's questions.What these assistants appear to neglect is providing users with different ways of answering their questions in order to improve understanding and task performance on the part of the user.The present study investigated whether such variability in responses to questions makes a difference in people's performance on a task that requires use of those answers.17 middle and high school students were asked, as part of a course on infectious diseases, to develop a plan for creating a vaccine.They were given a list of 13 questions they could ask and receive answers to.Half of the students had only informational answers available to them and the other half could choose from informational answers, real-world analogies, goal-based or causal-factor answers.Results showed that students who were given the choice of answer type to their questions received scores on their vaccine development solutions that were twice as high as the students who received informational answers only.Results suggest that personal assistants, including those embedded in educational technology, could be made more effective by offering a variety of ways to answer questions rather than a single format.
Authors: John Leddo, Tianle Chen, Aditya Menachery, Jai Agarwal, Taruna Agarwal
Language: English
Open Access status: diamond
OpenAlex ID: W4205242719
FWCI: 0.145
Cites:
Cited by: 10
Related to: 10