Download PDF
Dr Mariam Parwaiz
BHSc, MBChB, MPH (Hons)
Public Health Medicine Registrar
Counties Manukau District Health Board
The growth of artificial intelligence (AI) and the ongoing automation
of work are features of our time, and medicine will be increasingly
impacted by these trends. AI in medicine generally means the uti-
lisation of computer algorithms and automated processes to aid in
the diagnosis and treatment of patients. 1 The medical AI industry is
growing rapidly, and there has been an explosion of academic interest
in the subject. 2 Although medicine constantly evolves and adapts to
new technologies over time, health care systems tend to be naturally
risk-averse, and there is some caution within the medical community
about the role for AI in health care. 1 The medical students of today
will likely experience the opportunities and challenges associated
with AI in medicine throughout their future careers as doctors.
The role for medical education in New Zealand is to equip medical
students with the scientific knowledge and the professional skills and
attributes necessary to function effectively as doctors, and help them
progress towards mastering the science and art of medicine. While
AI in medical education is still a nascent idea in medical schools in
New Zealand, as time goes by and as AI inevitably becomes more of
a feature of medical practice, there will be a mounting pragmatic ne-
cessity for doctors, and for medical education, to engage with it. This
engagement should be done in an ethically-sound way, with the aim
of providing high-quality, equitable, culturally-safe, and patient-centric
care, in a manner that reflects the values and aspirations of health
care delivery in New Zealand.
Researchers recognise the potential of AI in medicine to improve
health care delivery, and current literature suggests that AI-based
tools can be as effective and accurate as human clinicians. 1 AI com-
petence will become an important skill to add to the vast skillsets
possessed by doctors. But doctors will not only need to be comfort-
able using AI in their day-to-day work, they will also need to have
an understanding of the principles behind both AI generally and the
specific AI-based tools they will use, as well as the benefits and po-
tential biases and flaws of these. Essentially, the doctor of the future
will exist in a world where they will need to be competent at using
AI; the role for medical education is to ensure future doctors are
prepared for that world.
Currently, medical education arguably does not sufficiently prepare
future doctors for the impending AI revolution in health care. To do
so will require a transformational reform in medical education, where medical students are taught traditional biomedical sciences and com-
passionate communication, alongside the principles of AI. 3 As medical
education academics are starting to argue, medical schools needs
to shift from focusing on information acquisition to ‘an emphasis on
knowledge management and communication’. 4
Within the medical profession itself, there is positive news. Recently
the New Zealand Medical Association (NZMA) commented on the
proposed World Medical Association (WMA) Statement on Arti-
ficial or Augmented Intelligence in Medical Care, suggesting that AI
should be regarded by the medical community as a technological tool
that can be applied to improve the quality and efficiency of health
care delivery and education. 5 It is important that medical organisa-
tions contribute to the discourse around AI in medicine, and it should
be interesting to read and review the WMA’s statement once it has
been ratified and published. The NZMA also suggested that the clin-
ical impact of interventions related to AI should be subjected to high
standards of empirical evaluation, with the possibility of unintended
negative consequences kept in mind and beneficial impacts not pre-
sumed. 5 This is a reasonable concern, and any AI-based tool devel-
oped should be robustly tested and validated before being deemed
suitable for wider general use.
In medicine it is always necessary to act in an ethical manner, and
with this in mind the Royal Australian and New Zealand College of
Radiologists recently produced a draft on Ethical Principles for AI in
Medicine. 6 The eight draft principles they identified, which will likely
be retained in the final version of the document, were: safety; avoid-
ance of bias; transparency and explainability; privacy and protection
of data; decision making on diagnosis and treatment; liability for de-
cisions made; application of human values; and governance. 6 These
principles, which are also relevant to AI as applied to other medical
specialities, provide an excellent framework to help ensure that AI in
medicine is, and continues to be, safe and effective.
Health equity is an important concern that must be kept foremost
in mind as AI is further adopted into medical practice. According
to the Ministry of Health, ‘In Aotearoa New Zealand, people have
differences in health that are not only avoidable but unfair and unjust.
Equity recognises different people with different levels of advantage
require different approaches and resources to get equitable health
outcomes’. 7 It is possible that introducing AI-based tools could have the unwelcome effect of increasing inequities between populations,
such as between socioeconomic groups, ethnic groups, or geographic
groups of people. For instance, the data that AI systems use could be
biased. 8 Data containing implicit ethnic, gender, or other biases will
generate results that are also biased. 8,9 Khullar provided a clear ex-
ample of this in a recent opinion piece, noting that if poorer patients
do worse after organ transplantation, AI algorithms may conclude
that such patients are less likely to benefit from treatment and thus
recommend against it, without accounting for or mitigating for wider
factors. 10 As medical professionals and custodians of the health care
system, we must ensure that technological advances in health care
are implemented systematically, are culturally safe and free from im-
plicit bias, and take account of the most vulnerable. We must ensure
that incoming AI tools do not, and will not, increase health inequities,
and preferably actually work to reduce the inequities we currently
see in health.
It is necessary to state that machines cannot and should not replace
human doctors. The role of the doctor will inevitably evolve over
time, but doctors will not become obsolete. Humans will always be
required to interpret outputs from machines, assess ethical and val-
ue-based dilemmas, and communicate empathetically. 11 The thera-
peutic relationship between doctor and patient is a fundamental tenet
of medicine and will remain so. There is no substitute for the human
touch. As AI becomes more and more a part of medical practice, the
role of medical education in imparting the soft skills of medicine will
increase in importance. These skills include an appreciation of ethics,
leadership skills, communication skills, and the ability to work in an
empathetic manner. 11 These skills are essential to being a good doctor,
and will continue to differentiate us from machines. 12 Hopefully we
can look forward to a future where AI tools work in an ethical and
equity-enhancing manner to complement our role as doctors and
improve our effectiveness in the health care system.
References
1. Loh E. Medicine and the rise of the robots: a qualitative review
of recent advances of artificial intelligence in health. BMJ Leader.
2018;2:59–63.
2. Kolachalama VB, Garg PS. Machine learning and medical education.
NPJ Digit Med. 2018;1:54.
3. Wartman SA, Combs CD. Medical education must move from
the information age to the age of artificial intelligence. Acad Med.
2018;93(8):1107–9.
4. Wartman SA, Combs CD. Reimagining medical education in the
age of AI. AMA J Ethics. 2019;21(2):E146–52.
5. New Zealand Medical Association. NZMA submission on
proposed WMA statement on artificial or augmented intelligence
in medical care [Internet]. 2018 [cited 15 Apr 2019]. Available at:
https://www.nzma.org.nz/__data/assets/pdf_file/0004/86818/
NZMA-Submission-on-proposed-WMA-statement-on-artificial-or-
augmented-intelligence-in-medical-care.pdf
6. Royal Australian and New Zealand College of Radiologists.
RANZCR ethical principles for AI in medicine – consultation
[Internet]. 2019 [cited 15 Apr 2019]. Available at: https://www.ranzcr.
com/our-work/advocacy/position-statements-and-submissions/
ranzcr-ethical-principles-for-ai-in-medicine-consultation
7. Ministry of Health. Achieving equity [Internet]. 2019 [cited 15 Apr
2019]. Available at: https://www.health.govt.nz/about-ministry/what-
we-do/work-programme-2018/achieving-equity
8. IBM. AI and bias [Internet]. 2019 [cited 4 Apr 2019]. Available at:
https://www.research.ibm.com/5-in-5/ai-and-bias/
9. Angwin J, Larson J, Mattu S, Kirchner L. Machine bias [Internet].
2016 [cited 4 Jun 2019]. Available at: https://www.propublica.org/
article/machine-bias-risk-assessments-in-criminal-sentencing
10. Khullar D. A.I. could worsen health disparities [Internet]. 2019
[cited 4 Jun 2019]. Available at: https://www.nytimes.com/2019/01/31/
opinion/ai-bias-healthcare.html
11. Balthazar P. Training medical students and residents for the AI
future [Internet]. 2018 [cited 14 Apr 2019]. Available at: https://www.
acrdsi.org/Blog/Medical-schools-must-prepare-trainees
12. Lauer AK, Lauer DA. The good doctor: more than medical
knowledge & surgical skill. Ann Eye Sci 2017;2(36).
Dr Mariam Parwaiz
BHSc, MBChB, MPH (Hons)
Public Health Medicine Registrar
Counties Manukau District Health Board
The growth of artificial intelligence (AI) and the ongoing automation
of work are features of our time, and medicine will be increasingly
impacted by these trends. AI in medicine generally means the uti-
lisation of computer algorithms and automated processes to aid in
the diagnosis and treatment of patients. 1 The medical AI industry is
growing rapidly, and there has been an explosion of academic interest
in the subject. 2 Although medicine constantly evolves and adapts to
new technologies over time, health care systems tend to be naturally
risk-averse, and there is some caution within the medical community
about the role for AI in health care. 1 The medical students of today
will likely experience the opportunities and challenges associated
with AI in medicine throughout their future careers as doctors.
The role for medical education in New Zealand is to equip medical
students with the scientific knowledge and the professional skills and
attributes necessary to function effectively as doctors, and help them
progress towards mastering the science and art of medicine. While
AI in medical education is still a nascent idea in medical schools in
New Zealand, as time goes by and as AI inevitably becomes more of
a feature of medical practice, there will be a mounting pragmatic ne-
cessity for doctors, and for medical education, to engage with it. This
engagement should be done in an ethically-sound way, with the aim
of providing high-quality, equitable, culturally-safe, and patient-centric
care, in a manner that reflects the values and aspirations of health
care delivery in New Zealand.
Researchers recognise the potential of AI in medicine to improve
health care delivery, and current literature suggests that AI-based
tools can be as effective and accurate as human clinicians. 1 AI com-
petence will become an important skill to add to the vast skillsets
possessed by doctors. But doctors will not only need to be comfort-
able using AI in their day-to-day work, they will also need to have
an understanding of the principles behind both AI generally and the
specific AI-based tools they will use, as well as the benefits and po-
tential biases and flaws of these. Essentially, the doctor of the future
will exist in a world where they will need to be competent at using
AI; the role for medical education is to ensure future doctors are
prepared for that world.
Currently, medical education arguably does not sufficiently prepare
future doctors for the impending AI revolution in health care. To do
so will require a transformational reform in medical education, where medical students are taught traditional biomedical sciences and com-
passionate communication, alongside the principles of AI. 3 As medical
education academics are starting to argue, medical schools needs
to shift from focusing on information acquisition to ‘an emphasis on
knowledge management and communication’. 4
Within the medical profession itself, there is positive news. Recently
the New Zealand Medical Association (NZMA) commented on the
proposed World Medical Association (WMA) Statement on Arti-
ficial or Augmented Intelligence in Medical Care, suggesting that AI
should be regarded by the medical community as a technological tool
that can be applied to improve the quality and efficiency of health
care delivery and education. 5 It is important that medical organisa-
tions contribute to the discourse around AI in medicine, and it should
be interesting to read and review the WMA’s statement once it has
been ratified and published. The NZMA also suggested that the clin-
ical impact of interventions related to AI should be subjected to high
standards of empirical evaluation, with the possibility of unintended
negative consequences kept in mind and beneficial impacts not pre-
sumed. 5 This is a reasonable concern, and any AI-based tool devel-
oped should be robustly tested and validated before being deemed
suitable for wider general use.
In medicine it is always necessary to act in an ethical manner, and
with this in mind the Royal Australian and New Zealand College of
Radiologists recently produced a draft on Ethical Principles for AI in
Medicine. 6 The eight draft principles they identified, which will likely
be retained in the final version of the document, were: safety; avoid-
ance of bias; transparency and explainability; privacy and protection
of data; decision making on diagnosis and treatment; liability for de-
cisions made; application of human values; and governance. 6 These
principles, which are also relevant to AI as applied to other medical
specialities, provide an excellent framework to help ensure that AI in
medicine is, and continues to be, safe and effective.
Health equity is an important concern that must be kept foremost
in mind as AI is further adopted into medical practice. According
to the Ministry of Health, ‘In Aotearoa New Zealand, people have
differences in health that are not only avoidable but unfair and unjust.
Equity recognises different people with different levels of advantage
require different approaches and resources to get equitable health
outcomes’. 7 It is possible that introducing AI-based tools could have the unwelcome effect of increasing inequities between populations,
such as between socioeconomic groups, ethnic groups, or geographic
groups of people. For instance, the data that AI systems use could be
biased. 8 Data containing implicit ethnic, gender, or other biases will
generate results that are also biased. 8,9 Khullar provided a clear ex-
ample of this in a recent opinion piece, noting that if poorer patients
do worse after organ transplantation, AI algorithms may conclude
that such patients are less likely to benefit from treatment and thus
recommend against it, without accounting for or mitigating for wider
factors. 10 As medical professionals and custodians of the health care
system, we must ensure that technological advances in health care
are implemented systematically, are culturally safe and free from im-
plicit bias, and take account of the most vulnerable. We must ensure
that incoming AI tools do not, and will not, increase health inequities,
and preferably actually work to reduce the inequities we currently
see in health.
It is necessary to state that machines cannot and should not replace
human doctors. The role of the doctor will inevitably evolve over
time, but doctors will not become obsolete. Humans will always be
required to interpret outputs from machines, assess ethical and val-
ue-based dilemmas, and communicate empathetically. 11 The thera-
peutic relationship between doctor and patient is a fundamental tenet
of medicine and will remain so. There is no substitute for the human
touch. As AI becomes more and more a part of medical practice, the
role of medical education in imparting the soft skills of medicine will
increase in importance. These skills include an appreciation of ethics,
leadership skills, communication skills, and the ability to work in an
empathetic manner. 11 These skills are essential to being a good doctor,
and will continue to differentiate us from machines. 12 Hopefully we
can look forward to a future where AI tools work in an ethical and
equity-enhancing manner to complement our role as doctors and
improve our effectiveness in the health care system.
References
1. Loh E. Medicine and the rise of the robots: a qualitative review
of recent advances of artificial intelligence in health. BMJ Leader.
2018;2:59–63.
2. Kolachalama VB, Garg PS. Machine learning and medical education.
NPJ Digit Med. 2018;1:54.
3. Wartman SA, Combs CD. Medical education must move from
the information age to the age of artificial intelligence. Acad Med.
2018;93(8):1107–9.
4. Wartman SA, Combs CD. Reimagining medical education in the
age of AI. AMA J Ethics. 2019;21(2):E146–52.
5. New Zealand Medical Association. NZMA submission on
proposed WMA statement on artificial or augmented intelligence
in medical care [Internet]. 2018 [cited 15 Apr 2019]. Available at:
https://www.nzma.org.nz/__data/assets/pdf_file/0004/86818/
NZMA-Submission-on-proposed-WMA-statement-on-artificial-or-
augmented-intelligence-in-medical-care.pdf
6. Royal Australian and New Zealand College of Radiologists.
RANZCR ethical principles for AI in medicine – consultation
[Internet]. 2019 [cited 15 Apr 2019]. Available at: https://www.ranzcr.
com/our-work/advocacy/position-statements-and-submissions/
ranzcr-ethical-principles-for-ai-in-medicine-consultation
7. Ministry of Health. Achieving equity [Internet]. 2019 [cited 15 Apr
2019]. Available at: https://www.health.govt.nz/about-ministry/what-
we-do/work-programme-2018/achieving-equity
8. IBM. AI and bias [Internet]. 2019 [cited 4 Apr 2019]. Available at:
https://www.research.ibm.com/5-in-5/ai-and-bias/
9. Angwin J, Larson J, Mattu S, Kirchner L. Machine bias [Internet].
2016 [cited 4 Jun 2019]. Available at: https://www.propublica.org/
article/machine-bias-risk-assessments-in-criminal-sentencing
10. Khullar D. A.I. could worsen health disparities [Internet]. 2019
[cited 4 Jun 2019]. Available at: https://www.nytimes.com/2019/01/31/
opinion/ai-bias-healthcare.html
11. Balthazar P. Training medical students and residents for the AI
future [Internet]. 2018 [cited 14 Apr 2019]. Available at: https://www.
acrdsi.org/Blog/Medical-schools-must-prepare-trainees
12. Lauer AK, Lauer DA. The good doctor: more than medical
knowledge & surgical skill. Ann Eye Sci 2017;2(36).