industry & reform
Dr Barbara Barbosa Neves studies the role of AI and robotics in relation to older people , connectedness and ageing . Picture : Supplied .
Understanding AI
World-first study links tech design to ageism .
By Elise Hartevelt
An award-winning sociologist is spreading awareness about how artificial intelligence ( AI ) could exacerbate ageism within the aged care sector .
Monash University ’ s Dr Barbara Barbosa Neves has received international recognition for her studies on ageing and socio-digital inequalities among older people .
Neves said the aged care industry should consider the potential impact of AI technology can have on the wellbeing , autonomy , and dignity of older people .
“ AI can perpetuate ageism and exacerbate existing social inequalities ,” she said .
“ When implementing AI technologies in aged care , we must consider them as
14 agedcareinsite . com . au part of a suite of care services and not as isolated solutions .”
Aged Care Insite spoke with Neves about how changing societal views on older people could positively impact the way AI developers create technology for aged care .
ACI : Can you give us a summary of your recent study ? BN : So the context for the study was that artificial intelligence technologies known as AI , and when we talk about AI , we ’ re talking about robots and voice assistance such as chat bots . These technologies are being pushed into aged care as a solution to the problems that are affecting the sector , including staff shortages . I guess we really wanted to understand the potential impacts of these technologies , including on how older people are viewed in both the design and in the implementation of AI . So a lot of the research that we have on AI from a social standpoint tends to focus on , for example , gender and racial biases . But the study of age related biases in AI is still quite new , particularly in relation to older people . And so this was a context for our research . We were trying to understand how older people were viewed in the design , but also implementation of AI .
And this is important because first of all , we assume that AI is more objective than humans in decision making processes . But what we know is that in reality , algorithms are based on human data and human values . And when we think about human data , we are thinking about datasets that use social behaviour and language and human categories to train AI models . And we know that those data sets are often biased because they tend to exclude particular populations . So they don ’ t fully include representative groups or representative data on particular groups . But in addition to how the data sets are designed and how we use data sets , we