From the hand of the University of California, Amazon It is driving the development of a voice assistant designed for older adults. To comply with this, the electronic market giant has awarded a group of computer scientists from this study house to work on a system that is capable of “understanding and answering” the questions of elderly people.
The minimum age of 65 has been set, because they are the people who tend to abandon voice assistants. Simply because their handling is complicated or they do not receive adequate responses from an AI. This is practically what happens to grandparents with other technologies, which they consider very advanced and end up rejecting.
Voice assistants process short requests
It is likely that when interacting with a voice assistant you will not present any problem because you have adapted to this technology, as it advances, you do too. However, when dealing with older adults, the situation changes, since they tend to ask long and complex questions for a system that is capable of processing short requests.
For example, when we make a request to the Apple voice assistant we say: “Siri do this or that”, we try to be as explicit as possible. In contrast, older people are used to “talking to people, not machines.”
Faced with this situation, Khalil Mrini, Ph.D. in computer science from UC San Diego, indicates that his “job is to close that gap… We are working on a technology that will change the load: instead of older adults changing or rephrasing their question, we want AI to be able to learn and understand the elderly ”.
As what is wanted is that the AI processes medical questions, Mrini proposes a response system that transforms from one extreme to another. This means reducing user questions to “key components.”
In this way they will be able to relate to a certain question of the more than 17 thousand medical questions from the National Institutes of Health. In this way, virtual assistants will be able to offer a clear and concise answer to their users.
This project is expected to help improve the way AI interprets human language. At the same time it makes it more accessible and inclusive.
How does AI behave when interacting with humans?