AI Regulation, Robots and Employment

'Models of Law and Regulation for AI' by Nicolas Petit and Jerome De Cooman comments 

This paper discusses models of law and regulation of Artificial Intelligence (“AI”). The discussion focuses on four models: the black letter model, the emergent model, the ethical model, and the risk regulation model. All four models currently inform, individually or jointly, integrally or partially, consciously or unconsciously, law and regulatory reform towards AI. We describe each model’s strengths and weaknesses, discuss whether technological evolution deserves to be accompanied by existing or new laws, and propose a fifth model based on externalities with a moral twist.

'The Losers of automation: How the introduction of robotics changed the European occupational class structure' by Paolo Barbieri, Giorgio Cutuli and Saverio Minardi comments 

This paper examines the occupational and social impact of labour-replacing technologies in Western European countries. Our study combines data on job tasks created with O*Net 3.0, information on regional robots exposure created with data from the International Federation of Robotics, and microdata from the EU-LFS from 1997 to 2017 that were longitudinalised by applying pseudo-panel models. 

We investigate unemployment risks associated with the introduction of labour-saving technologies as well as changes in workforce composition and possible modifications to the occupational stratification of European societies caused by the introduction of these new technologies. In addition, we examine the “shrinking-middle-class” hypothesis as we are interested in analysing the impact of robotics on social stratification in Europe. 

Our results reveal that the degree to which the process of technological innovation is embedded in institutions is highly relevant to understanding the stratification effects of robotics. This finding calls into question the deterministic – that is, the intrinsically functionalistic – perspective that underlies the economic approach of SBTC/RBTC theories. Overall, we find that a process of technical change lies at the root of different upgrading scenarios in Europe, with Northern Europe demonstrating the most positive effects of this process. Central Europe appears more stable, and technology here currently seems to be inducing only mild growth in non-manual and non-routine jobs, a reduction in unemployment risks, and an overall “upgrading” of employment- and class structure. The “losers” of the process of technological change appear to be the Southern European countries, which are currently experiencing a reduction in employment levels for low- and mid-educated male workers as well as an overall downgrading of occupational and class structures. Overall, our results do not confirm any convergent trend of technological unemployment or of “hollowing out the middle class”

COVID Liability

'Products in a Pandemic: Liability for Medical Products and the Fight against COVID-19' by Duncan Fairgrieve, Peter Feldschreiber, Geraint Howells and Marcus Pilgerstorfer in (2020) 11(3) European Journal of Risk Regulation 565-603 comments 

A multitude of medical products are being developed and produced as part of efforts to tackle COVID-19. They are varied in nature and range from test kits to tracing apps, protective equipment, ventilators, medicines and, of course, vaccines. The design, testing and manufacture of many of these products differs from production in normal times due to the urgency of the situation and the rapid increase in demand created by the pandemic. This article considers the legal issues arising as a result of the production of emergency products, particularly from a products liability perspective. To what extent do existing concepts under the European Product Liability Directive, such as defect, causation and the various defences, permit the pandemic to be taken into account when a Court is considering issues of liability? What is the impact on liability of the modified regulatory regime? In light of that discussion, the case for alternative responses is examined from a comparative and European perspective, including the issue of Government indemnities for the manufacturers of products, legal exemptions from liability and alternative no-fault compensation schemes.

Mammoths

'Patentability and de-extinct animals in Europe: the patented woolly mammoth?' by Aisling McMahon and David M Doyle in (2020) Journal of Law and the Biosciences comments 

De-extinction is a hot topic within conservation science but the potential patentability of de-extinct animals in Europe has hitherto been unexplored. This article addresses this lacuna, examining the legal, commercial, and ethical implications of patenting de-extinct animals under European patent law. The article is organized into four parts. Part I explores the reasons why patents are relevant and may be applied for in this context. Part II provides an overview of the scientific techniques currently being used in de-extinction projects, setting the foundation for the analysis of patentability which follows. Part III then critically assesses whether recreated animals would qualify as patent eligible subject matter under European patent law. It also investigates the extent to which European patent exclusions such as those on animal varieties, essentially biological processes, and the morality provisions might apply and whether recreated animals would meet the novelty requirement for patentability. Part IV concludes by highlighting the possible ramifications of patenting such animals, elucidating the chasm between the cultural and symbolic significance held by such animals, and their lack of differential treatment in the patent law sphere. It argues that de-extinction reignites questions around the scope of patents, and the role of ethical considerations within patent decision-making which warrant urgent reconsideration. 

 The authors argue 

 Ever since the 1993 blockbuster film ‘Jurassic Park,’ the notion of bringing vanished species back to life has hovered on the boundaries of reality and science fiction, but scientists are now reputedly on the brink of de-extinction. While conservationists argue that this risks persuading lay people that extinction is reversible, technological advances in synthetic biology and genetics have made the revival of extinct species a real scientific possibility, with much of the recent literature in this area shifting from whether we can resurrect extinct species to whether we should. In May 2016, for instance, the International Union for the Conservation of Nature Species Survival Commission released ‘Guiding Principles for Creating Proxies of Extinct Species for Conservation Benefit’ (2016), while scientists, conservationists and bioethicists, inter alia, have begun to consider the ethics of reviving extinct species, the optimal candidate species for de-extinction, as well as attempting to conduct cost-benefit analyses of both resurrection and reintroduction. The ‘nascent discipline’ of de-extinction has been the topic of animated debates within the scientific community,5 but at least one point of consensus has emerged: we are now on the brink of de-extinction. As Donlan puts it, ‘[t] here are many unknowns surrounding de-extinction. Whether it will happen or not, however, is likely not to be one of them.’ 

Much of the academic literature to date has understandably focused on the science of bringing back extinct species, but ‘de-extinction presents us with myriad ethical, legal and regulatory questions.’ These include a range of questions around the legal status of de-extinct animals, including potential legal obligations to protect them, and the resulting environmental issues which may arise. Moreover, an important but underexplored question is the extent to which such animals, if successfully recreated, could be commercialized. Watching certain resurrected species, especially ‘cool’ and ‘charismatic megafauna,’ is expected to be an exciting diversion for many (indeed, it is often considered, albeit anecdotally, to be a central motivation for de-extinction), but little academic work has focused on the ‘non-ecological “instrumental values” that de-extinct species would be likely to have for human beings, most notably, their commercial value.’ One avenue to obtain such value from de-extinct animals is via the patent system, and specifically applying for a patent on the revived animal, yet it remains an open question whether such animals would be patentable in the European context, if ongoing de-extinction attempts are successful. In taking this focus, the article explores the legal, commercial, and ethical implications of ‘patenting’ de-extinct animal species. 

This article breaks new ground by exploring the potential European patent law implications of de-extinction in relation to animal life. In doing so, this paper focuses on animal life, as opposed to plant or other organisms, which could also be recreated via de-extinction. The rationale underlying this approach is rooted in the fact that it is these projects (e.g. particularly the Woolly Mammoth project) that have captured the public imagination and provoked the most scholarly concern to date. For this reason, we wish to explore how applications for patents on such animals in Europe may play out if such megafauna are successfully recreated. 

In examining this issue, the article focuses on the European patent law context for two main reasons: first, to date, academic inquiries of the legal implications of de-extinction have focused almost exclusively on domestic law in the USA. This is not to suggest that these studies are less valuable for adopting such an approach—the USA, after all, is one of the traditional centers of conservation power and where key de-extinction projects are ongoing but many of the species that are considered viable candidates for some form of de-extinction are not species whose habitats or migration routes are exclusive to, or even include, the USA. The applicable laws outside the US context are in vital need of investigation if we are to preemptively address the legal issues that could conceivably arise should de-extinction become feasible in the next few years. To date, these aspects of species recreation and particularly the patent implications in Europe have remained unexplored. This article aims to address this lacuna by furnishing a detailed analysis of de-extinction under European patent law. 

Secondly, the focus on European patent law was taken because Europe is one of the few patent jurisdictions with express moral exclusions from patentability. Therefore, examining the European context provides a useful site to examine the ethical ramifications of granting patents in the de-extinction context and the extent to which such ethical issues are likely to be considered within patent law. It explores how the current exclusions under European law—which have proved so controversial in the past for transgenic animals—could potentially be interpreted in the de-extinction context. Given that European patent law has express moral exclusions from patentability, concerns posed by patents on de-extinct animals could theoretically (at least on paper) have a greater chance of consideration within the European context. However, the article will argue that given past interpretative practices surrounding such provisions in Europe, without institutional change within patent law, such ethical considerations are unlikely to be significant hurdles to patenting such technology in Europe (contingent, of course, on how the science develops). 

This analysis has important practical and theoretical significance. From a practical perspective, the article highlights the main patentability questions within European law that will arise in light of de-extinction projects, aiming to provide an important reference point within the de-extinction debate. At a theoretical level, it identifies aspects of European biotechnological patent law that will require further clarification should de-extinction come to fruition, such as definitional questions surrounding the patentability of cloned animals and transgenic animals not created for medical research purposes. As will be demonstrated, these issues also have broader significance for advanced biotechnologies more generally. 

In conducting this analysis, the article is structured as follows: Part I demonstrates why patents are relevant and could be applied for in the context of de-extinction projects; Part II then briefly maps the scientific avenues that are being used in de-extinction projects setting the foundation for the legal analysis which follows; Part III provides a detailed assessment of the potential patentability of de-extinct animals, examining questions of patent eligible subject matter; the potential applications of the exclusions from patentability under Art 53 EPC to animals created in the de-extinction context; and whether such animals would meet the patent novelty requirement. Part IV concludes by highlighting the possible effects, and positive/negative ramifications, of patenting such animals and elucidates the chasm between the cultural and symbolic significance held by such animals, and their lack of differential treatment in the patent law sphere.

Audit

The Australian National Audit Office has an inconvenient habit, consistent with its statutory charter, of politely questtioning Government ineptitude or downright disregard of law. That's resulted in the funding restrictions that are evident in other agencies: if you can't abolish it, just make sure it's kept on short rations. 

The Parliamentary Joint Committee of Public Accounts and Audit is now conducting an Inquiry into the Auditor-General Act 1997 (Cth), reflecting ss 8(g) and 8(h) of the Public Accounts and Audit Act 1951 (Cth). 

The terms of reference are 

 To inquire into and report on the adequacy of general provisions contained in the Auditor-General Act 1997 (the Act), with particular reference to:

• the governance framework as it relates to the Auditor-General and the Australian National Audit Office (ANAO), including the independence of the Auditor-General as an Officer of the Parliament and the audit independence of the ANAO, and resourcing arrangements; 

• the Auditor-General’s information gathering powers and confidentiality of information, including with reference to parliamentary privilege and the interaction between the Freedom of Information Act 1982 and the Act; 

• the interaction of the Act and other relevant legislation including the Public Governance, Performance and Accountability Act 2013, the Public Accounts and Audit Committee Act 1951, Freedom of Information Act 1982, and the Parliamentary Privileges Act 1987; 

• the Auditor-General’s capacity to initiate audits into, and examine the performance of all entities in the Australian Government sector; accessibility and transparency of reports and audit conclusions, including the operation of section 37 of the Act; 

• the Audit Priorities of the Parliament; 

• the role and appointment of the Independent Auditor; and any related matters. 

Deaths

Having made a somewhat acerbic submission to the Senate Foreign Affairs, Defence and Trade Committee regarding the two National Commissioner for Defence & Veteran Suicide Prevention Bills I was interested to see today's release by the Australian Institute of Health and Welfare (AIHW) of its third annual update of information around suicide among serving, reserve and contemporary ex-serving Australian Defence Force (ADF) personnel. 

The report is stated as serving to 'inform improvements in suicide prevention and other services for serving and ex-serving ADF personnel and their families', reflecting ongoing failure at the Department of Veterans' Affairs noted by the Productivity Commission last year and a succession of parliamentary committees. 

It indicates that men who were currently serving full-time or in the reserve were considerably less likely to die by suicide than Australian men generally. That was not the case for ex-serving men, who were 21% more likely to die by suicide than their counterparts in the general community after adjusting for age. Although e the rate of suicide for ex-serving women is lower than that for ex-serving men, ex-serving women were more than twice as likely to die by suicide than other Australian women after adjusting for age. Men who were discharged from the ADF for medical reasons were more likely to die by suicide than those discharged for voluntary reasons. For ex-serving men, the rate of suicide was 28 per 100,000, which was higher than the rates for serving (11 per 100,000) and reserve men (12 per 100,000), with 465 suicide deaths among serving, reserve and contemporary) ex-serving ADF personnel between 2001 and 2018.

Recusal

In Andrew Learmont v SAS Trustee Corporation [2020] NSWDC 595 Neilson DCJ states 

This is an application that I "recuse" myself. The use of such terminology is deplorable. The Latin root is the verb, recuso, recusare, recusavi, recusatum. That is a technical term in Roman law. It is used by certain jurists meaning, to make an objection or to demur. It is so used by Cicero, Celsus, Quintilian and Julian. It gives a noun form, recusatio, recusationis, which means either a counter plea or a demurrer. However, the verb could also be used generally, meaning to make an objection or to protest. It is so used by Cicero, by Ovid, by Seneca, by Livy and by Caesar. It can also mean, not to accept or consent, or to decline, to reject, or to oppose. It is not used reflexively and is not used in Latin to mean to disqualify oneself. 

According to the second edition of the Oxford English Dictionary published in 1989, the word, recuse, is rare and means, to refuse or to make an objection. There is a form of the word, recusal, but that means an objection to a judge as being prejudiced, but not the action of the judge in disqualifying himself. However, the primary meaning assigned to one form of the word, recuse, was that of being a “recusant” and the associated status of recusancy. A recusant was a person who refused to attend the services of the Church of England as established by Parliament. It applied to all who refused to attend the services of the Church of England but, in particular, members of the Catholic faith. The second edition of the Oxford English Dictionary does not admit the use of the verb "recuse" to mean, the action of a judge in disqualifying himself. 

The 5th ed of the Shorter Oxford English Dictionary published in 2002, does give the verb, recuse, a fourth meaning, when used reflexively, of a judge withdrawing from a hearing of a case because of a possible conflict of interest or lack of impartiality. However, it is clear, from the other entries in the Shorter Oxford English Dictionary that that use is United States usage. It does not admit to the usage being either British, Australian, Canadian, New Zealander or South African. 

However, the verb, recuse, in English has the same meanings as it does in Latin, to refuse a thing offered, to reject or renounce a person or his authority or to object or to refuse to do something. The use of the word "recuse" being a request of a judge to disqualify himself from the hearing of a case is not consistent with its Latin etymology, is inconsistent with English usage and, although it may be used in the United States of America, it is not part of the Queen's English and its use is to be eschewed.

I take the application to be one that I disqualify myself. I accede to that application.

Vaping

The Senate has established a Select Committee on Tobacco Harm Reduction, the latest in a series of parliamentary and other inquiries.  

The Select Committee will inquire (quickly) into tobacco reduction strategies, with particular reference to:

• the treatment of nicotine vaping products (electronic cigarettes and smokeless tobacco) in developed countries similar to Australia (such as the United Kingdom, New Zealand, the European Union and United States), including but not limited to legislative and regulatory frameworks; 
• the impact nicotine vaping products have had on smoking rates in these countries, and the aggregate population health impacts of these changes in nicotine consumption;
• the established evidence on the effectiveness of e-cigarettes as a smoking cessation treatment; 
• the established evidence on the uptake of e-cigarettes amongst non- smokers and the potential gateway effect onto traditional tobacco products; 
• evidence of the impact of legalising nicotine vaping products on youth smoking and vaping rates and measures that Australia could adopt to minimise youth smoking and vaping; 
• access to e-cigarette products under Australia’s current regulatory frameworks; 
• tobacco industry involvement in the selling and marketing of e-cigarettes; and 
• any other related matter.

The 2018 House of Representatives Standing Committee on Health, Aged Care and Sport report on Use and Marketing of Electronic Cigarettes and Personal Vaporisers in Australia responded to terms of reference requiring it to 

inquire into and report on the use and marketing of electronic cigarettes (E-cigarettes) and personal vaporisers in Australia, in particular: 

 

1. The use and marketing of E-cigarettes and personal vaporisers to assist people to quit smoking; 

 

2. The health impacts of the use of E-cigarettes and personal vaporisers; 

 

3. International approaches to legislating and regulating the use of E-cigarettes and personal vaporisers; 

 

4. The appropriate regulatory framework for E-cigarettes and personal vaporisers in Australia; and 

 

5. Any other related matter. 

It followed the 2017 report by the Senate Community Affairs Legislation Committee regarding the Vaporised Nicotine Products Bill 2017.

The 2018 House of Reps report - where members were in strong disagreement - offered the following recommendations - 

1   that the National Health and Medical Research Council fund an independent and comprehensive review of the evidence relating to the health impacts of electronic cigarettes (E-cigarettes). This review should be updated every two years to take into account the findings of new research into E-cigarettes. Topics covered by the review should include:

• The effectiveness of E-cigarettes as an aid to help people quit smoking tobacco cigarettes; 

• The health effects of ingredients commonly used in E-cigarette liquids. Following the review, any ingredients found to have significant negative impacts on human health should be prohibited from use in E-cigarette liquids; 

• The likelihood that E-cigarettes will increase the number of young people using nicotine and the number of young people smoking; 

• The health impacts of long term E-cigarette use; 

• The relative health impacts of E-cigarettes as compared to tobacco products.

2   that the Department of Health convenes an international meeting of health experts from similar economic jurisdictions to discuss different policy and legislative approaches to electronic cigarettes.  

 

3   a national approach be taken to the regulation of non-nicotine electronic cigarettes. 

 

4   that the Therapeutic Goods Administration continues to oversee the classification of nicotine and relevant exemptions, and the assessment of any electronic cigarette product as a therapeutic good. 

 

5   that the Australian Government establish a regulatory process for assessing and, if necessary, restricting colourings and flavourings used in electronic cigarettes.

Last month's ANU 'Summary report on use of e-cigarettes and relation to tobacco smoking uptake and cessation, relevant to the Australian context' by Emily Banks, Katie Beckwith and Grace Joshy (National Centre for Epidemiology and Population Health) for the Australian Department of Health) draws on the following reports: 

1. Review of smoking prevalence and trends in Australia 2. Review of smoking prevalence and trends in the Aboriginal and Torres Strait Islander population 3. Review of evidence regarding changes in smoking behaviour with decreasing smoking prevalence, including that relating to the “hardening hypothesis” 4. Review of the patterns of e-cigarettes use 5. Systematic review of evidence regarding combustible smoking uptake in relation to e-cigarette use; and 6. Systematic review of evidence regarding the efficacy of e-cigarettes for combustible tobacco or nicotine cessation. 

 The authors state 

 Despite world-leading tobacco control, smoking remains Australia’s leading cause of preventable disease and death, including for Aboriginal and Torres Strait Islander peoples. In many countries, e-cigarettes are explicitly or implicitly marketed as aids to smoking cessation and, among e-cigarette users, smoking cessation is a commonly reported reason for use. However, the substantial majority of smokers who quit successfully do so unaided and no e-cigarette products have been approved by the Australian Therapeutic Goods Administration as smoking cessation aids; the situation is similar in many other countries. In the context of the exceptional harms of combustible tobacco use and the current promotion of e-cigarettes it is important to review the current relevant evidence on e-cigarettes and smoking behaviour, to support informed decision-making.

The key findings are - 

 Smoking prevalence and trends in Australia 

• From the 2019 National Drug Strategy Household Survey, 11.0% of people aged 14 and over in Australia were current daily smokers (equivalent to 2.3M people): a statistically significant drop from 12.2% in 2016, and following sustained reductions in smoking prevalence over recent decades. These falls were largely driven by younger people not taking up smoking. 

• The proportion of never smokers has increased over time, in adults and particularly among youth in Australia. 96.6% of youth aged 14-17 in 2019 had never smoked, with youth never-smoking prevalence increasing five-fold between 2001 and 2016. 

• The majority of Aboriginal and Torres Strait Islander people do not smoke. The prevalence of current daily smoking was 40.2% in 2018/19 in Aboriginal and Torres Strait Islander adults, following substantial and significant reductions over the past decade, particularly in major cities and regional areas and among younger adults and among youth. 

“Softening” of the Australian smoking population over time 

• Based on published reviews and large-scale repeat cross-sectional studies from Australia and similar high-income countries, declining smoking prevalence has generally been accompanied by increasing motivation to quit, reduced dependency and greater quit rates among smokers. 

• In 2010, an estimated 2.0% of the general population of Australia aged 18 and over were smokers who were unmotivated to quit and had difficulty quitting. 

• Based on the weight of the available evidence from Australian and similar high-income countries, the “hardening hypothesis” – proposing increasing difficulty quitting among the population of smokers as smoking prevalence declines – can be rejected. This concept should be replaced by the evidence- based conclusion that declining smoking prevalence is accompanied by “softening” of the smoking population, whereby smokers are, on average, more readily able to quit. 

Patterns of use of e-cigarettes 

• The prevalence of e-cigarette use varies widely between countries and has increased substantially in many countries over the past decade, particularly among young people. In countries where e- cigarettes are available as consumer goods, such as the United States, use is becoming common, particularly among youth, with recent data indicating 10-20% of US high school children report recent use of e-cigarettes. 

• Among people in Australia aged 14 years and over in 2019, 11% had ever used e-cigarettes, most of whom (60%) reported using e-cigarettes once or twice only; 2.0% (equivalent to 412,000 people) reported current use (daily, weekly or monthly) and 1.1% (equivalent to 227,000 people) reported daily use, according to the 2019 National Drug Strategy Household Survey. Use has increased significantly over the last six years. 

• Among people in Australia aged 14 years and over in 2016, 0.5% were estimated to be dual current users of e-cigarettes and combustible cigarettes, and 0.2% dual daily users; current e-cigarette use was reported by 4.4% of current combustible cigarette smokers, 1.5% using e-cigarettes daily. 

• Population patterns consistent with short-term use of e-cigarettes for smoking cessation would predominantly include past use of e-cigarettes in ex-smokers, little ongoing use in current smokers and virtually no use in never-smokers. 

• However, among current daily e-cigarettes users in 2016, 32% were also daily smokers, 11% were non-daily smokers, 38% were ex-smokers and 18% were never-smokers. Hence, an estimated 43% of daily e-cigarette users in Australia in 2016 were dual e-cigarette users and combustible tobacco smokers. In 2019, 1.9% of ex-smokers reported past use of e-cigarettes and 5.2% of never-smokers reported ever using e-cigarettes. 

E-cigarette use in never- and former-smokers, and combustible tobacco smoking uptake 

• Observational evidence from three systematic reviews and 25 primary research studies was included. 

• A meta-analysis of data from these studies showed that never smokers who have used e-cigarettes were, on average, around three times as likely as those who have not used e-cigarettes to try smoking conventional cigarettes and transition to regular tobacco smoking. All studies found evidence of an increased risk, with wide variation in the magnitude of this risk. 

• Where evidence on nicotine content was available, it indicated that a substantial majority of e- cigarettes in these studies delivered nicotine. 

• The studies reviewed were observational in nature as it is not ethical or appropriate to randomise non-smokers to e-cigarette exposure. The quality of the evidence from the reviews and primary research studies was rated moderate overall. 

• The limited available evidence indicates that former smokers who had used e-cigarettes were around twice as likely to relapse and resume current smoking as those who had not used e-cigarettes. 

The efficacy of nicotine-delivering e-cigarettes as an aid to smoking or nicotine cessation 

• Reliable evidence on the efficacy of use of e-cigarettes for smoking cessation requires large-scale, independent randomised controlled trial (RCT) evidence from multiple studies. 

• The randomised controlled trials identified found no significant difference in quit rates between smokers randomised to nicotine-delivering e-cigarettes versus no intervention or non-nicotine e- cigarettes, although point estimates of risk ratios were all above 1.00. 

• Of the four trials of nicotine-delivering e-cigarettes versus other nicotine-replacement therapy identified, three found no statistically significant difference in smoking cessation between the groups and one found significantly greater cessation with nicotine-containing e-cigarettes than with other nicotine-replacement therapy. 

• Data derived from three RCTs suggest that smokers randomised to nicotine-containing e-cigarettes versus other types of nicotine-replacement therapy were substantially more likely to be using any form of nicotine (i.e. e-cigarettes or nicotine-replacement therapy) at follow up. One study found that around 80% of successful quitters randomised to e-cigarettes continued to use them at one- year follow up while 9.0% of those randomised to other nicotine-replacement continued to use it. 

• The overall quality of the evidence was rated as low and uncertain: the few RCTs conducted were generally small, employed a wide range of study designs across diverse settings and the majority had methodological issues indicating a high risk of bias. 

• Overall, there is insufficient evidence that nicotine-delivering e-cigarettes are efficacious for smoking cessation, compared to no intervention, placebo existing nicotine-replacement therapy or other best-practice interventions. However, preliminary evidence highlights the potential for nicotine-delivering e-cigarettes to support cessation, and more reliable, large-scale evidence is needed.