Please use this identifier to cite or link to this item:
http://hdl.handle.net/11375/29002
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Emerson, Claudia | - |
dc.contributor.author | Alievska, Belinda | - |
dc.date.accessioned | 2023-10-04T19:36:41Z | - |
dc.date.available | 2023-10-04T19:36:41Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | http://hdl.handle.net/11375/29002 | - |
dc.description.abstract | As artificial intelligence (AI) increasingly revolutionizes healthcare, the development of AI-powered medical tools holds promise for improving patient care by increasing the accuracy of diagnoses, providing precise treatments, and more. Despite the promise of AI tools, they suffer from a critical limitation due to a lack of representation of elderly experiences. In this paper, I examine the multilayered issue of the underrepresentation of the elderly in the data used in the development of medical AI tools, assessing its implications and proposing relevant solutions. This thesis is divided into five chapters, with Chapter One providing an overview of the current state of AI, highlighting the potential of AI-powered tools for improving healthcare outcomes, and their relevant limitations. Chapters Two to Four assess how various sources of AI training data (i.e., clinical trials, electronic health records [EHRs] and self-reporting tools) fail to adequately represent elderly experiences. Clinical trials have long suffered from a lack of elderly representation, originating from efforts to protect vulnerable populations from research-induced harm. While the attempt has been to protect, the result has been to arbitrarily exclude older adults from participation in clinical research. As such, there is a lack of diverse data that accurately reflects the complexities of their unique medical needs. Further, the presence of age-based bias towards the elderly presents an additional layer of concern. The manifestation of such biases in EHRs perpetuates inaccurate data that informs the development of medical AI tools, consequently maintaining disparities in healthcare. Additionally, self-reporting tools fail to account for the distinct cognitive and physical abilities of older adults, presenting useability challenges that result in the inadequate representation of elderly experiences in the data. Finally, the paper culminates with a compilation of proposed solutions to address the underrepresentation of elderly experiences in medical AI development. These solutions propose efforts to improve elderly participation in clinical trials, efforts to mitigate physician bias in EHRs, and the design of self-reporting tools that are cognizant of the unique needs of older adults. | en_US |
dc.language.iso | en | en_US |
dc.title | Ethical Challenges Arising from the Underrepresentation of the Elderly in the Development of Artificial Intelligence (AI) for Medical Applications: Exclusion, Bias, and the Limits of Accessibility | en_US |
dc.type | Thesis | en_US |
dc.contributor.department | Philosophy | en_US |
dc.description.degreetype | Thesis | en_US |
dc.description.degree | Master of Arts (MA) | en_US |
Appears in Collections: | Open Access Dissertations and Theses |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Alievska_Belinda_2023September_MA.pdf | 753.6 kB | Adobe PDF | View/Open |
Items in MacSphere are protected by copyright, with all rights reserved, unless otherwise indicated.