Governance of artificial intelligence and personal health information
| DOI | https://doi.org/10.1108/DPRG-08-2018-0048 |
| Pages | 280-290 |
| Date | 13 May 2019 |
| Published date | 13 May 2019 |
| Author | Jenifer Sunrise Winter,Elizabeth Davidson |
| Subject Matter | Information & knowledge management,Information management & governance,Information policy |
Governance of artificial intelligence
and personal health information
Jenifer Sunrise Winter and Elizabeth Davidson
Abstract
Purpose –This paper aims to assess the increasing challenges to governing the personal health
information (PHI)essential for advancing artificial intelligence(AI) machine learning innovations in health
care. Risksto privacy and justice/equity are discussed,along with potential solutions.
Design/methodology/approach –This conceptual paper highlights the scale and scope of PHI data
consumedby deep learning algorithms and theiropacity as novel challenges to health data governance.
Findings –This paperargues that these characteristicsof machine learning will overwhelmexisting data
governance approaches such as privacy regulation and informed consent. Enhanced governance
techniques and toolswill be required to help preserve the autonomy and rights of individuals to control
theirPHI. Debate among all stakeholdersand informed critiqueof how, and for whom, PHI-fueledhealth AI
are developedand deployed areneeded to channel these innovationsin societallybeneficial directions.
Social implications –Health data may be used to address pressing societal concerns, such as
operational and system-level improvement, and innovations such as personalizedmedicine. This paper
informs work seekingto harness these resources for societal good amidstmany competing value claims
and substantialrisks for privacy and security.
Originality/value –This is the first paper focusing on health data governancein relation to AI/machine
learning.
Keywords Big data, Governance, Artificial intelligence, Deep learning,Personal health information
Paper type Conceptual paper
1. Introduction and motivation
Artificial intelligence (AI) technologies increasingly enable innovations from searching the
internet to voice and facial recognition, smart appliances, and even to driverless cars. In the
past, key limitations of AI have been the availability of sufficient data for training algorithms
and the inability of AI systems to manage data in their natural form. Now, with omnipresent
digitalization of data about humans and their activities, deep learning[1] algorithms
increasingly are able to take advantage of stockpiles of “big data” to enhance a learning
model’s performance and extendthe sophistication and reach of AI applications (Chen and
Lin, 2014;Jordan and Mitchell, 2015).
AI innovations are particularly promising in the domain of health and health-care services.
From personalized health care tailored for each individual’s biology to improvements in
health-care delivery systems, AI innovations are projected to revolutionize health-care
outcomes for individuals and for health-care systems (Flores et al., 2013). Vital to these
potential innovations are the vast stockpiles of individual-level health data needed for deep
learning models. Now, personal health information (PHI)[2] data stores, such as notes from
routine visits to the doctor, medical imaging, self-monitoring of steps, sleep and heartbeats,
and DNA repositories, are rapidly accumulating, and will over time (given much-needed
improvements in data quality and standardization), be applied to train deep learning
algorithms in the growing array of AI health-careapplications (Miotto et al.,2017).
Jenifer Sunrise Winter is
Associate Professor at the
School of Communications,
University of Hawaii at
Manoa, Honolulu, Hawaii,
USA. Elizabeth Davidson is
Professor at the Shidler
College of Business,
University of Hawaii at
Manoa, Honolulu, Hawaii,
USA.
Received 30 August 2018
Revised 16 December 2018
Accepted 17 December 2018
PAGE 280 jDIGITAL POLICY, REGULATION AND GOVERNANCE jVOL. 21 NO. 3 2019, pp. 280-290, ©Emerald Publishing Limited, ISSN 2398-5038 DOI 10.1108/DPRG-08-2018-0048
To continue reading
Request your trial