Forum Recap: ‘Regulating Future Technologies’
On Monday, 20 November, the Research For Legal Change held its second forum: 'Regulating Future Technologies'. Two Bristol postgraduate students, Louise Hatherall, and Kit Fotheringham, were invited to share their research.
Louise's research, 'Exploring The Impact of Granting Gene Patents on Human Health, Competition and Innovation', delved into genetic patents and their impact on research, health and human society.
Inspired by Moore v Regents of California (1990), where a patient treated by a doctor had their DNA extracted without their knowledge nor permission and the doctor went on to make billions off the patent, Louise sought a definite way to define the extent to which we own our DNA. The EU and US have different legal regulations; the EU has legislated via its European Biotechnology Directive that anyone can claim a fragment of DNA, so long as they state their intent. Under US Law, the Supreme Court coined the term 'anything man-made under the sun' for non patentable items, which is vague but would seemingly disqualify all DNA.
Louise shared the detrimental impact of patentable DNA, where patent owners have a twenty year monopoly over their fields. Patenting recipes for chocolate bars would be an industrial advantage, but patenting treatments and DNA is more morally dubious. For example, the BRCA mutation is responsible for an 80% increased risk in breast cancer. The mutation itself, was patented by a big Pharma, Myraid. Research showed that 12% of data was missing, but the monopoly prevented any others from doing research on it. Myraid, free from competition, could charge exorbitantly - $3000 per test. This adversely affected the decisions of many individuals and their families, especially since it was a requirement to get tested for breast cancer in the US when applying for employment, else the individuals would go discriminated against if untested. Additionally, there were only 2 centres for BRCA research in the entire US, making the vital service even less accessible.
Central to Louise's research was AMP (The Association for Molecular Pathology) v Myraid Genetics Inc (2013). The AMP challenged that Myraid's test was not the most cost effective, and that they were hiking up the prices and taking unfair advantage of their market monopoly. While Myraid had unrivaled profits and hence were able to fund their legal costs, the American Civil Liberties Union jumped in, arguing with a class action lawsuit that Myraid's restriction of tests through price amounted to a civil rights issue, and the ACLU won. This marks an optimistic stronghold that Public Interest Groups have against big pharmas, even though they are often ignored or sidelined.
The rest of the industry and legislation has a longer way to go towards a more equitable system. The CRISPR Gene Editing tool, a nanobot DNA editing tool, is only licensed to two universities, restricting access, application and research. Ancestry is a biobank corporation, which traces clients' linage via DNA samples sent in - however, they own the DNA that they receive, which is compiled and sold to big pharmas.
Louise's research involved Literature Reviews, Case Studies and Interviews, and the challenging aspect of her paper is that new data and cases are always emerging and even in the time that she has began her research, the field has changed radically, so she is expecting to rewrite her paper once all her data is collated and finished.
Kit's research, on the other hand, deals with data configuration at public and private sectors by AI. Contrary to some popular belief, academia and legislation are rising up to meet the new demands of this emerging technology.
There are four types of Artificial Intelligences. Type 1's can react to stimulus, with a program knowledge of rules, but can only work within a fixed environment and react to pre-programmed stimuli; for example, a chess-playing AI. Type 2's can find solutions to problems within its environment, but only has limited awareness of that environment; for example, driverless cars, which do not possess total knowledge of their environment, but can use ingenuity in coding to change their reactions, like speeding up and slowing down. Type 3's and 4's do not yet exist, but are in the pipeline; Type 3's have a representation of the world around it within itself, called a Theory Of Mind. Type 4's are self-aware AIs, true consciousnesses, with their own feelings, moods, tastes, opinions and self-grown values.
The problem that already exists with Type 1 and Type 2 Artifical Intelligences is that they use Machine Learning to quickly analyse vast sets of data, taking their own initiative to produce the optimal solution on their own. However, the data is collected by humans, who are innately biased on their own. In this light, the optimisation process would merely serve to reinforce the biases already present.
In light of these huge developments, governments have surprisingly little resources to deal with them, as compared to a large MNC or a centralised government. Small governments like Enfield City Council in London and Birmingham City Council already use machine learning AIs in customer service, but they typically do not have much to invest in technology, which culminates in lower access to it and software which is not entirely customisable to their needs.
Of course, legislation exists to protect the greater interests of the masses. The Data Protection Act 1998 controls how personal data can be used, and the upcoming Gene Protection Regulations Act of May 2018 makes it possible for individuals to request from companies all the data they have on the individual, but it is still free from obligation to declare how that personal data is being used. It would become meaningless to use the acts as a data control mechanism because ultimately, our personal data is still being used and abused. There is also a random element to Machine Learning, which sometimes produces meaningless code from data. Legislation also takes a while to come into force, after which the situation may have changed radically.
A possible solution is more transparency between the individual and the corporation that holds their data. However, open source has not always worked well. It led to the infamous Heartbleed Bug in 2016, being circulated in open source code for two years without detection, before finally wrecking a significant amount of damage when triggered. Tort and Judicial Review, according to Kit, go out the window simply because many cases are decided on policy. The saving grace is that the Supreme Court still has a say in the most important cases that help to improve the policy over time, to strike a better balance.
The Research Forum for Legal Change is most sincerely thankful to Louise and Kit for their time and brilliant presentations offering us a glimpse into how future technology has to be regulated and the means of doing so. Do connect with us via our social media to find out about our Forums as they come!