Ethical issues in use of Big Data and Artificial Intelligence for IVF treatment

Angelica Cheng

Active Member
https://www.todayonline.com/commentary/commentary-deeper-discussion-needed-over-ethical-issues-use-big-data-and-ai-ivf-treatment-2117396

Commentary: Deeper discussion needed over ethical issues in use of big data and AI for IVF treatment

In recent years, the development and application of new reproductive technologies have attracted much attention in Singapore, as a means of overcoming the country’s severe demographic challenges.

An exciting development is the novel application of big data and artificial intelligence in IVF (in vitro fertilisation) treatment.

Indeed, ethical issues relating to the application of these new technologies in human biomedical research are anticipated to become a hot topic for debate in coming years, which is why the Bioethics Advisory Committee (BAC) in Singapore is studying this issue closely, and lists this as one of its current projects. A public consultation may thus be in the offing.

The BAC was set up in 2000 to specifically address ethical, legal and social issues arising from biomedical research, and make relevant policy recommendations to the Singapore Government.

DIRE NEED FOR BETTER IVF PROCEDURES

In the field of fertility treatment, this is anticipated to yield better diagnostic procedures, as well as improved and streamlined IVF protocols for patients.

There is a dire need for better IVF procedures to overcome age-related decline in fertility due to late marriages and delayed parenthood, which is a major contributory factor to Singapore’s dismal birthrates.

For example, big data and artificial intelligence can be used to predict and optimise an individual patient’s bodily response to varying dosages of different fertility drugs based on medical condition, to maximise chances of success in IVF treatment, in what is referred to as “precision fertility medicine”.

In a recent scientific breakthrough, artificial intelligence and machine learning can also be used to screen microscopy images of IVF embryos, to predict their risks of genetic abnormalities and developmental potential to give rise to a healthy baby.

While such advancements in medical technology would certainly be a boon to IVF patients, there is also much potential for abuse.

In particular, sex selection and genetic testing of IVF embryos for non-disease socially-desirable traits such as good looks and intelligence.

In recent years, Singapore has emerged as a hub for collating big data on the genomics of various Asian ethnic populations for application in drug discovery and precision medicine, which is facilitated by the country’s diverse multi-racial population originating from different parts of Asia.

Such big data could thus provide a fertile ground for the development of new artificial intelligence-based screening technologies.

TWO KEY AREAS OF ETHICAL CONCERN

Here, two key areas of ethical concern pertaining to the application of big data and artificial intelligence in fertility treatment will therefore be highlighted.

The first relates to the development of artificial intelligence (AI) algorithms to assess the genetic normality of embryos based only on microscopy images, with some degree of accuracy.

This is particularly useful for older women undergoing fertility treatment, who are at much higher risks of having genetic abnormalities in their IVF embryos, as compared to younger women.

For example, the “Life Whisperer Genetics” AI algorithm developed by the American health care company Presegen is non-invasive, low-cost, and provides results instantaneously, thereby making it preferable to the tedious and time-consuming pre-implantation genetic testing (PGT) of IVF embryos.

According to Presagen Chief Medical Science Officer Dr Sonya Diakiw, although this AI screening technique, based only on microscopy images alone, may not be as accurate as manual genetic testing itself that involves actual DNA sequencing; a relatively high accuracy rate of 77.4 per cent can still be achieved, as reported in the reputable medical journal “Human Reproduction”.

This still has much room for further improvement. Perhaps in the future, higher accuracy may be achieved by combining AI screening of microscopy images with other parameters such as embryo secretion profile.

Although there is nothing ethically or morally controversial in evaluating the genetic normality of IVF embryos based on microscopy images, it must be noted that the very same AI algorithm can easily be modified to predict and select the sex of IVF embryos.

This would certainly be a cause of legal and ethical concern, because non-medical sex selection of IVF embryos is banned by health regulations in Singapore[X9] .
Such a practice is widely condemned to be immoral and unethical as it reinforces traditional gender bias and sexism, and encourages couples to favour having children of one gender over the other - eliciting much controversy in many Asian countries where women are subjected to strong societal discrimination and gender-based violence.

MORE VIGILANCE NEEDED

Hence it is imperative that the Singapore Ministry of Health should exercise vigilance on the application of such AI-based screening technologies on IVF embryos, to ensure that these are not being abused for illicit sex selection.

The other area of major ethical concern is the potential abuse of big data and artificial intelligence in the genetic testing and selection of IVF embryos for non-disease socially-desirable traits.

These include intelligence, athletic prowess and physical traits related to beauty standards such as tallness, fair complexion, hair and eye colour, in what is known as preimplantation genetic testing for polygenic risks (PGT-P).

As such socially-desirable characteristics are complex traits determined by the combination of multiple genes, polygenic risk scores (PRS) are used to estimate an individual embryo’s likelihood of developing an adult-onset, multifactorial trait by analysing the combination of specific genetic variants within its genome.
This thus represents a fertile ground for the application of artificial intelligence on big data.

There are minimal risks involved here, as there is no genetic modification, and it is basically a technique for picking the “winning ticket” in the “genetic lottery” for intelligence, good looks, and athletic prowess.

This is a particularly lucrative business model, because parents naturally and instinctively want the best for their children, particularly in a hyper-competitive society such as Singapore, where the educational rat-race often starts early in life.

Indeed, a recent large-scale survey conducted in the USA showed that the desire to use genetic testing to have more intelligent children is not confined to idiosyncratic and fringe elements within society, with 38 per cent of respondents saying that they would do so if given the chance.

More worryingly, 28 per cent of respondents said they are even willing to resort to gene editing to achieve this goal.

Such figures might even be higher in Singapore, given the prevalent “kiasu” (afraid to lose) mentality in the country.

Nevertheless, it must be noted that there is an important distinction between embryo testing and selection to avoid serious harm from known genetic defects and for so-called ‘enhancement’, such as good looks, higher IQ and greater athletic prowess.

The latter raises a range of ethical concerns and is a highly controversial issue that has provoked much-heated debate among healthcare professionals, biomedical scientists, lawmakers and religious leaders.

Hence, besides banning local clinical applications of AI-based embryo screening technologies for selection of gender and non-disease socially-desirable traits, the Government should also enact corresponding parallel legislation on big data collection locally in Singapore for the development of such unethical applications overseas.

An outright ban on local data collection for unethical applications overseas may be difficult to legislate and enforce.

It would be rather challenging to prosecute violators in more liberal jurisdictions where such applications are legal, given the cross-jurisdictional aspect in proving that local data had indeed been used abroad.

This is an area that will need to be further studied and debated by various relevant bodies such as the BAC, in consultation with the public.
 
  1. What are some concerns that the Singapore government might have when it comes to using AI in IVF treatment?

    The primary concern here is the potential abuse of AI to screen and select gender and non-disease socially-desirable traits such as intelligence, good looks and athletic ability. The willingness to use new technologies to have genetically-enhanced children is not confined to idiosyncratic and fringe elements in society, as shown by a recent large-scale survey conducted in the USA. Out of 6,800 people polled, 38% of respondents said that they would use genetic testing of IVF embryos to improve their children’s chance of entering an elite university. More disturbingly, some 28% of respondents stated that they are even willing to use gene editing to achieve this goal. Such figures may likely be higher in Singapore, given the pervasive “kiasu” afraid to lose mentality of Singaporeans.

  2. Pros/Cons of using big data/AI in genetic enhancement of IVF embryos

    On the Pro side, there is the concept of procreative beneficience, by which would-be parents have an inherent right and moral responsibility to beget children who are most likely to experience the greatest well-being – that is, the most advantaged child, the child with the best chance at having the best life.
    On the Con side, is the possibility of worsening social equality, because only the rich can afford to use such expensive technologies to beget genetically-advantaged children.


  3. You say it’s possible to abuse AI in IVF treatment in illicit sex selection. How does this work?

    AI technology can be used to track IVF embryo development from microscopy images taken at various time points, and predict the likelihood of that embryo being genetically abnormal. Such a technology can easily be tweaked instead to also predict the sex of the embryo.
    Here, there there could be a slippery slope situation. For example, in some countries, PGT or manual genetic testing of IVF embryos is often marketed as a means of avoiding Down syndrome in older women. But, on the sidelines, patients also so sex selection because genetic testing inevitably also reveal the sex of embryos. Hence, a slippery slope could also develop for AI, whereby the need to prevent Down syndrome is used as a handy excuse for sex selection.


  4. If there is demand to use AI/big data to select IVF embryos based on traits we prefer (aka genetic testing), how likely will it go ahead here in Singapore?

    Very unlikely, given that Singaporean society as a whole is rather conservative. Not only the general public, but also the medical community and politicians as well. Very likely, AI technologies will be approved only for preventing genetic diseases in IVF treatment.
    The ban on genetic enhancement with AI technology will instead lead to medical tourism, whereby Singaporean couples wanting genetically-enhanced children travel overseas to do the procedure.


  5. Using AI and machine learning to check on the health of both mummy and embryo sounds like a good thing. But how necessary is it?

    Highly necessary, given the rising incidence of late motherhood in Singapore, and increased risks of older women giving birth to a child with genetic defects, most commonly Down syndrome. Unlike conventional prenatal testing that takes place after pregnancy, abortion can be avoided by screening IVF embryos before transferring them to the would-be mother.

  6. One of these AI treatments has an accuracy rate of only 77.4 per cent. Is that good enough especially if we’re dealing with life?

    This is just the beginning. As time progresses, the technology will surely mature and higher accuracies will definitely be attained. The advantages of AI screening of embryos, are it’s much lower costs and faster speed compared with manual preimplantation genetic testing. Moreover, it is non-invasive whereas manual genetic testing involves extracting cells from the embryo, which risks causing damage.

  7. What other ways can we study age-related decline in fertility?

    In the context of overcoming poorer quality eggs and sperm in older patients, the solution will be freezing eggs and sperm at a younger age. More recently, there is also the emerging technology of ovary tissue freezing for younger women, which can be used to restore fertility as well as overcome menopause in older women.
 

Back
Top