THE DESIGN AND TESTING OF DIGITAL SYSTEMS NEEDS TO BE RIGOROUS TO ENSURE THAT THESE INFERENCES ARE LIMITED AND UNDERSTOOD AS FAR AS POSSIBLE . be done with any data and an easy way to opt out .
E D I T O R ' S Q U E S T I O N
NEIL MASON
PRINCIPAL CONSULTANT , METHODS ANALYTICS
When implementing new technology or digital strategies in healthcare , a number of ethical considerations come into play .
Firstly , safety , which may be regulated when it applies to medical devices , and increasingly to AI , but due to the wide reach of data , may be more difficult to assure than ever before .
A second issue is fairness , across a number of domains , for example , with digitally enabled services , are all segments of the community equally and sufficiently able to engage with a service , perhaps through their smartphones ? In addition , are the technologies biased , either through their design or the data that may have been used to train or build part of the system ?
Privacy is a third consideration , and one of particular concern as we become ever more digital . The trend is for technologies to collect as much information as possible , however , systems employed in healthcare should be careful to collect only what is necessary for a purpose , and this collection and processing needs to be underpinned by clear communication of what will
THE DESIGN AND TESTING OF DIGITAL SYSTEMS NEEDS TO BE RIGOROUS TO ENSURE THAT THESE INFERENCES ARE LIMITED AND UNDERSTOOD AS FAR AS POSSIBLE . be done with any data and an easy way to opt out .
A serious ethical question is how to enable people to access a service who are unable to or do not wish to consent to such data collection . A new concern in the age of data science is where a system may be able to infer characteristics of people based on other data points . The design and testing of digital systems needs to be rigorous to ensure that these inferences are limited and understood as far as possible . Finally , in regard to privacy , security is a major consideration , and the use of suitably tested secure environments and protocols is crucial .
Additionally , transparency is a key ethical consideration today , particularly as some digital systems now act as a black box , with few people able to explain how a technology comes to the conclusions it does . As patients and service users become more tech-savvy , clear explanations need to be in place to build trust .
Frameworks to help consulting professionals include legislation such as GDPR , which functions as a useful reminder of a number of ethical principles as well . In addition , when applying data science and machine learning , frameworks such as that designed by The Institute for Ethical AI and Machine Learning offer applicable steps to ensure advancements bring benefits to services and users in an ethical way and without bringing harm .
26 www . intelligenthealth . tech