Aisling de Paor highlights some of the ethical and legal concerns arising with use of algorithms based on genetic information, and advocates the need to appropriately control and regulate these new technologies.
Technological advances and a growing use of algorithms in commercial, medical and social contexts are pushing the boundaries of existing law and policy and provoking a range of concerns. There have also been rapid advances in genetic science, revealing a vast amount of detail about an individual's current and future health, predisposition to disease, disability and potential behavioural traits. Genetic science has significant application in medical contexts, where it promises to transform the delivery of health care, and also in social decision-making processes where it may be used as a powerful predictive tool.
In addition to use for therapeutic purposes, genetic and other health information is increasingly being used (and will inevitably be used in the future) in commercial settings such as insurance, employment and banking. Use of genetic information in algorithmic decision-making processes is generally motivated by economic and efficiency factors, where the predictive nature of genetic information is particularly attractive. For example, employers may use genetic-based algorithms as a predictive tool in the recruitment process, to select the most genetically desirable employees, those perceived to be most productive, and those unlikely to become ill in the future. Similarly, insurance companies may use such algorithms to select the most low-risk and profitable customers.
The growing use and application of these algorithms which process genetic information create concerns from the perspective of individual rights and raises the spectre of unintended consequences.
Technological advances and genetic information
Genetic science and technological innovation is evolving rapidly. It has the potential to revolutionise health care and our understanding of the genetic make-up of human beings, with more effective diagnosis, treatment and even eradication of many diseases. Genetic advances have led to unprecedented scientific progress in the understanding of disease and genes, and have facilitated a growth in the practice of genetic testing. The predictive nature of genetic information is important in what it potentially reveals about future predisposition to disease, disability and behavioural and personality tendencies.
Technological innovation has also propelled the development and application of algorithms as a selection and efficiency tool. The convergence of big data, genetics and new technologies pushes us toward a new future in decision-making processes. Genetic-based predictive analytics offers efficiency and has significant application, particularly in commercial settings including employment and insurance, where it may be used as an economically motivated predictive tool. For example, in insurance, genetic-based algorithms may be viewed as an optimisation mechanism that will encourage enhanced risk management.
What are the concerns?
Technological advances and the growing use of algorithms create new economic opportunities, provoking a range of ethical, legal and other concerns. These advances highlight a conflict of competing rights between commercial interests and fundamental rights. Commercial actors such as employers and insurers may wish to access and use genetic information for economic reasons, to select or retain the most genetically desirable individuals – in other words, the most productive, healthy employees or the most profitable insurance applicants. In circumstances where genetic information is used to predict future health and behaviour in automated decision-making processes, particular concerns arise regarding an individual's right to privacy and the right not to be discriminated against.
Unauthorised access to and use of genetic information may violate an individual's genetic privacy. Considering the uniquely private and sensitive nature of genetic information, there are compelling reasons to protect the right to privacy in these circumstances. Considering the familial nature of genetic information (which expands the proportion of individuals vulnerable to privacy violations), there are further incentives to acknowledge and protect privacy rights. In the USA, the 2015 landmark case of the 'devious defecator' robustly affirmed the need to protect genetic privacy and found a violation of the federal Genetic Discrimination Non Discrimination Act 2008 (Lowe and Reynolds v. Atlas Logistics Group Retail Services LLC, U.S. District Court for the Northern District of Georgia, No. 13-cv-2425). In this case, two employees were wrongly accused of defecating in the employer's warehouse and, on foot of this, they were subsequently compelled by their employer to take a genetic test. They successfully claimed that their genetic privacy had been breached, resulting in subsequent violation of their right to dignity and integrity. There is a strong individual and societal interest in ensuring the protection of privacy, as reflected by the strong data protection laws in the EU. In particular, the recently introduced GDPR provides specific privacy protections for genetic data, creating obligations for those involved in processing genetic data.
The right not to be discriminated against on the grounds of genetic information may also be threatened through use of genetic-based algorithms. By introducing a new layer of human difference, the concept of 'genetic discrimination' describes discrimination on the basis of genetic information, genetic make-up or family medical history. Decision-making on the basis of the probability of an individual developing a disease or disability (either through manual decision making or automated decision making) may be deemed to be unlawful discrimination. Discrimination on the basis of genetic information is arguably as unjust as discrimination based on gender or race, or other characteristics over which an individual has no control. Although there is no specific European level legislation prohibiting genetic discrimination, there is strong legislation in Europe providing for equality in the workplace (under the Employment Equality Directive 2000), on a number of grounds including disability. These protections impose obligations on employers to avoid discrimination. In addition, the EU Charter of Fundamental Rights has highlighted genetic discrimination as a priority area of reform in the EU, signalling future legislative developments in this area.
The right to dignity and integrity is also at stake in these circumstances. Respect for dignity is undermined or denied in circumstances where persons with particular genetic predispositions are discriminated against, devalued or their lives disregarded in decision-making processes. This may have undesirable societal implications by categorising individuals according to perceived genetic status. In addition, the existence of barriers to social goods and services (from misusing genetic information) may result in the exclusion of certain vulnerable individuals (those with disabilities and genetic predispositions to disease) and may further violate principles of dignity and integrity.
Conclusion: wider public policy and other concerns
The misuse or the unregulated use of genetic information and genetic-based algorithms may ultimately lead to the creation of a genetic underclass in society. This genetic underclass and the potential relegation of genetically undesirable individuals may arise as a result of the misapplication of new technologies, and as a result of the decisions yielded from use of algorithms.
In addition to privacy and discrimination concerns, the uncontrolled use of genetic-based algorithms may also exacerbate societal stigma (towards disability, disease and difference), and highlight prejudice in society. <pq2>If a certain belief, or misperception is introduced into a data collection or decision-making system (for example, in an employment recruitment process), this may lead to more biased and prejudiced algorithms. Such (structural) bias may amplify concerns of discrimination, stigmatisation and other undesirable implications of misusing genetic information.</pq2>
The particular nature of genetic information is also a cause for concern with use of genetic-based algorithms. Problems may arise regarding knowledge of genetics including the predictive nature of genetic information and what it can reveal. In consideration of the rapid pace of progress in this field, it is important that those using genetic-based algorithms have sufficient knowledge of the nature and limitations of genetics. Linked to this concern, it is important to note the limited scientific accuracy (or predictive value) of most genetic tests at the moment (creating further potential for misinterpretation of genetic information), with particular reference to the interaction of genes and environment factors in disease expression. In the future, it is likely however that these technologies will become more refined and accurate in terms of predictive value.
In addition to law and policy intervention, scientific and technological developments need to be monitored. In light of the limited predictive value of most genetic information, there is a need to ensure human input, supervision and oversight in use of genetic-based algorithms. In consideration of the unique nature of genetics, and the need for appropriate knowledge of the field, genetic expertise is particularly important to avoid misinterpretation, unintended consequences, misuse and bias. There is a need for training and awareness-raising for those using genetic information in decision making processes. Similarly, transparency regarding the manner in which these new technologies are used is also important, to enhance public confidence in science and new technologies.
Aisling de Paor is a Lecturer in Law at the School of Law and Government, Dublin City University and a solicitor.