IEEE Consumer Electronics Magazine - July/August 2023 - 18

Towards Sustainable AI in Modern Medicine
the FDA's current approach to software modifications,
the FDA anticipates that many of these artificial
intelligence and machine learning-driven
software changes to a device may need a premarket
review. " The US FDA published an action plan
in Jan 20218 for using AI in healthcare devices,
expressing the intent to establish " good machine
learning practices, " oversight of how algorithms
behave in real-world scenarios and development
of research methods for rooting out bias. However,
these intentions have not yet been integrated
into a formal regulatory framework. More
recently, in June 2022, the FDA published guidelines
on AI in radiological devices, requiring companies
to outline how the technology is supposed
to perform and provide evidence that it works as
intended. In September 2021, the European Commission
(EU) also published a proposal for a legal
framework as part of the AI Act, to promote the
development of an ecosystem of trust with AI.36
The legislation presented a risk-based regulatory
approach to AI, identifying healthcare as a highimpact
sector for AI and requiring AI-based
healthcare devices to fulfill the requirements of
the AI Act as well as those already set under the
existing EU Medical Device Regulation. These are
steps in the right direction. However, many gaps
exist across these regulatory frameworks.37 For
instance, the US guidelines do not stipulate rigorous
prerequisites and manufacturer responsibility
like the EU regulations do, for lifecycle regulation
and algorithmic bias mitigation in AI-based healthcare
devices. Conversely, the US guidelines call for
promoting greater transparency and publishing
summaries for each approved healthcare device,
which is not the case in EU where their database
on medical devices (Eudamed2) is not publicly
accessible.
To inspire long-term recognition of ethical principles,
formal recognition of AI development for
healthcare as a profession with equal standing as
for other high-risk professions may be necessary.
It is a regulatory anomaly that professions that
provide a public service are licensed, but the profession
responsible for designing technical systems
to augment and replace human decisionmaking
in public healthcare remains unlicensed.32
Finally, many AI ethics initiatives to date have created
professional codes of ethics that address
design goals and the behaviors and values of individuals
but ignore the validity of specific applications
and their underlying organizational and
business interests.38 This approach conveniently
steers debate towards the wrongdoings of unethical
individuals, and away from the overall failure
of business models and unethical organizations.
Regulations will be essential to hold erring businesses
accountable. Such regulations can also target
other big issues such as fair distribution of
healthcare benefits, protecting equality of care,
and promoting ethical societal values in the era of
AI-driven healthcare.
CONCLUSION
In this article, we reviewed the landscape of
ethical challenges facing AI integration into
healthcare products. AI introduces many new
ethical concerns for healthcare products related
to transparency, bias, responsibility, safety, privacy,
justice, and autonomy that go beyond traditional
medical ethics principles, and must be
addressed during the design, deployment, and
use of these systems. Broadly speaking, healthcare
systems in many ways act as the core of
modern societies. If ethical mistakes are made in
these early days of adoption and implementation
of AI in healthcare, the fallout could undermine
public trust, which could be devastating for the
healthcare industry and also cause patients to
look for their healthcare from outside of formal
systems where they may encounter significant
risks. To address this challenge, we
discussed multiple approaches that will be
crucial to promote the ethical integration of
AI in healthcare.
ACKNOWLEDGMENT
This work was supported by the National Science
Foundation under Grant CNS-2132385.
&REFERENCES
1. World Health Organization, " A healthier humanity: The
WHO investment case for 2019-2023, " No. WHO/DGO/
CRM/18.2, 2018.
18
IEEE Consumer Electronics Magazine
IEEE Consumer Electronics Magazine - July/August 2023

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - July/August 2023
