Monday, August 26, 2019
Carl Benedikt Frey's "The Technology Trap" (2019): A Holistic, Historical Examination of Artificial Intelligence and the Future of Work
A frequently-voiced social concern, universally, is the perceived growing threat of artificial intelligence (AI) to eliminate the jobs of millions of workers. The concern is voiced by workers’ compensation lawyers and others in the community in a more narrow, existential way. If the number of jobs is significantly truncated, particularly those in the industrial sector, will workers’ compensation become superannuated, and along with it those who labor in the dispute resolution process?
A healthy commentary exists in this realm. In a new book, Oxford University economic historian Carl Benedict Frey takes a retrospective/historical look at the situation and tries to predict the future from experiences of the past. See The Technology Trap: Capital, Labor, and Power in the Age of Automation (Princeton Univ. Press 2019).
Frey, a specialist in studying technology and employment, has, along with his colleague Mark Osborne, published several widely-read articles on this topic the last three years. The book is thick but, in the end, highly readable and a balanced, up-to-date treatment of this burning socio-legal issue. The book does not focus on industrial injuries. Still, the advance of AI, and the replacement of human labor with robots and/or other labor-saving innovations, holds the portent that many of the injuries and deaths encountered in the present day will be eliminated in the future.
An attractive feature of the book is the author’s attention to the workers whose jobs are threatened by AI. Will they simply accede, as they did with computerization, to the elimination of jobs, or will they seek government action to suppress AI innovation? Meanwhile, will workers be attracted more and more to populist movements, fired by resentment of elites who seem most advantaged by job-threatening technological processes?
A key focus of the book is explaining that the advance of technology has, since the eighteenth century – first in England and then in the United States – affected labor in two ways. The first affect is from technology that simply replaces jobs that were theretofore undertaken by human labor. Frey’s example, throughout, is that of gas lamplighters. With the invention of the light bulb, such jobs were simply eliminated. The second affect is from technology that, in contrast, enables workers to undertake the same or similar jobs with greater ease, and hence with greater productivity; and/or which generates new, theretofore uncontemplated, jobs. Frey again uses electrification as an example, here with the technological marvel of the lighting and powering of factories. Between lighting which allowed longer hours of work and the powering of motors to drive manufacturing processes, workers could be more productive, avoid dangerous conditions and, ultimately, achieve greater wages.
Of course, workers are better off when technology generates jobs that are of the enabling sort, as opposed to that which totally eliminates jobs. However, even with enabling technology, history has shown that a period of adjustment, which he calls “Engels’ Pause,” may exist. During such a period, workers may be displaced from their customary employments and experience depressed wages and other social and economic disruption. This phenomenon is most vividly illustrated by the Luddites, with their rebellion against mechanization and their destruction of labor-replacing machinery. Their revolt against innovation was not based on some abstract devotion to custom and tradition but, instead on the real threat of unemployment and other socio-economic upheaval.
On this point, Frey, throughout, makes an observation that may be counterintuitive. In this regard, over the centuries, rulers, particularly monarchs, often joined in the suppression of technological development because of social disruption. They feared that the population, particularly enterprises like craft guilds, would turn against them. In the late eighteenth and early nineteenth centuries, however, commercial interests in England gained formidable political clout and began to displace landed aristocratic interests. With this change in power structure, the government began supporting technological innovation. Indeed, the army was called out to crush the Luddites. Meanwhile, the new governing elites did nothing (at first) to prevent factory owners from replacing able-bodied men with women and children – who could easily, and cheaply, labor at the new machinery. In any event, the pattern over the last two centuries has been for government to support technological innovation in the workplace. The anxiety over displaced (and displeased) workers has persisted, but the typical response, with a few exceptions, has been retraining programs and unemployment compensation.
Frey thoroughly covers the past in trying to predict the AI future. Most familiar to the current reader will be how workers and society reacted to such things as the automation of automobile and other factories, the invention of the typewriter, and computerization of innumerable processes. By and large, these technologies have been ones of enablement, however hard for some the transition. In contrast, Frey believes that AI is full of replacement-of-jobs potentialities. He identifies, in particular, truck driving. The coming loss of jobs to autonomous vehicles in this area is especially critical, as truck driving is a leading form of employment in virtually all states. He ponders, as noted at the outset, whether those currently employed in the field, and in other entry-level and low-skilled jobs, will simply roll over in the face of the phenomenon. The author seems certain that at least some workers will suffer through another “Engel’s Pause” period of adjustment, such as was encountered so painfully during the initial phase of the industrial revolution.
The author’s final chapter is devoted to recommendations to make the period of adjustment easier. First and foremost is the promotion of education; it is the non-educated who, in the wake of IA innovations, will suffer the most. The second is retraining and a more flexible approach to educating the displaced. Frey also discusses wage insurance, tax credits (he is unsympathetic to universal basic income), decreased regulation, i.e., licensure, of skilled jobs (which he favors), relocation, and modifying transportation systems to connect displaced workers in one area to others where work is plentiful.
In the end, Frey has no hard answers to the labor issue in the age of automation, but a considered reading of his book is thought-provoking and, gratifyingly, places the issue in detailed historical context.
Wednesday, August 21, 2019
Researcher Frank Neuhauser Writes on Firefighter Cancer Presumptions: Observations from the Pennsylvania Trenches
In an important new article, distinguished California researcher Frank Neuhauser identifies the current trend of states enacting and expanding cancer presumptions in favor of firefighters. He cites and evaluates the studies which have sought to ascertain whether firefighters are at increased risk of developing cancer and – along with others, including IARC – finds that they are inadequate in evidencing such a phenomenon. At best prostate, testicular, and brain cancer may be implicated, but in Neuhauser’s view the science is weak even as to these types of cancer in establishing a pattern of causation. See Frank Neuhauser, Cancer Presumption for Firefighters: Good Policy or Give Away?, IAIABC Perspectives, p.7 et seq. (July 2019) (behind a pay-wall: https://www.iaiabc.org/iaiabc/Perspectives.asp).
Neuhauser unequivocally recommends that legislatures not enact such statutes. He declares, “What does all this mean for policymakers? The evidence for elevated risk of cancer among firefighters should be considered insufficient for adopting presumptions…. Presumption[s] lead to [unjustifiable] extra costs being imposed on local jurisdictions ….”
True, some studies have suggested increased risk, but Neuhauser finds three major weaknesses attendant to such studies. First, most firefighters have group health insurance, so cancers are more readily identifiable in that occupation than in the general population. Second, firefighters, overall, live longer than members of the general public and are hence more likely than others to develop cancer. Third, studies that show greater incidence are retrospective in nature and, for the most part, observed Caucasian men, a group that has a substantially greater rate of cancer than the general population.
Neuhauser also rejects the proposition that the cancer presumption in the instance of firefighters should be expanded to other maladies, such as heart disease, hypertension, Tuberculosis and other respiratory diseases, PTSD, and MRSA. “Again,” he admonishes, “the evidence for elevated risk among firefighters for any of these conditions is nonexistent, inconsistent or even contradictory.”
Neuhauser’s critique is familiar to the Pennsylvania reader. In the wake of Act 46, the scientific integrity, or lack of the same, of the firefighter cancer presumption has been well-aired in the court precedents. A minor library of cases is, indeed, available where Dr. Tee Guidotti, testifying for the City of Philadelphia, questions the presumption and ridicules the firefighters’ expert, a Dr. Singer, for suggesting that occupational cancer in a firefighter can be deemed exposure-related based not on epidemiological evidence but, instead, via the differential diagnosis method.
Of course, we know that physicians (and judges) cannot “attack” or “reject” a statutory presumption. That is a lesson from both the workers’ compensation and Black Lung litigation of many years back. Neuhauser knows this, and his remedy is that legislatures should never feel pressured into enacting such scientifically-bereft legal devices in the first place.
Readers of Neuhauser’s essay may assume that any firefighter who develops cancer and receives a presumption of causation automatically receives an award. But the notion that the cancer-victim firefighter is always on the receiving-end of the purported “Give Away” is incorrect. Presumption laws create rebuttable presumptions. In practice, municipalities often, if not always, oppose such claims and are often successful. In Pennsylvania, meanwhile, the presumption of causation is a mere procedural device, and when the municipality produces any level of contrary causation evidence, the burden of proof is once again on the firefighter to prove causation – just like in any other case. (In contrast, in some states, like Virginia, the firefighter’s cancer presumption is substantive in nature, that is, an element of proof, and aids the cancer victim throughout any dispute.)
In any event, given the procedural character of the presumption in Pennsylvania, some observers have questioned whether the presumption device really assists the disease-victim worker. Notably, during the dust disease wars of the 1970s and 1980s, injured workers (coal miners alleging pneumoconiosis and steelworkers alleging silicosis) never relied solely on the presumption. To the contrary, they always presented expert testimony in support of causation.
In arguing against presumptions, Neuhauser may have launched the proverbial “assault on a citadel.” He realizes that, in the present day, firefighters have a strong lobby and are politically powerful. However, presumptions, though currently trending, are not menacing 21st Century phenomena like legalized marijuana and comfort dogs on airplanes. To the contrary, presumptions have been found in workers’ compensation laws for over a century. Notably, a disease-of-the-heart-and-lungs presumption for firefighters has been a feature of Pennsylvania law since 1939.
Meanwhile, the British Workmen’s Compensation law of 1906 contained a list of diseases, to which attached presumptions of causation relative to workers who had exposure to the related disease hazards: “If the workman at or immediately before the date of the disablement … was employed in any process mentioned in the [list of occupations] … and the disease contracted is the disease … [paired with such occupations] … the disease, except where the certifying surgeon certifies that in his opinion the disease was not due to the nature of the employment, shall be deemed to have been due to the nature of that employment, unless the employer proves the contrary.” True, cancer in firefighters was not on the list – the law concerned itself more with anthrax in wool sorters and hookworm disease in coal miners. The point to remember, however, is that presumptions in workers’ compensation laws are not some modern device, custom-manufactured to coddle to the lobby of a special constituency.
Thursday, August 15, 2019
I am scrambling with intensity trying to convert a 5-credit, two-semester torts course to a 4-credit, one-semester torts course. Occasionally, the mind wanders . . .
A long, long time ago, in late 19th century Wisconsin, 12 year old Putney kicked 14 year old Vosberg in the leg within their classroom. Why? Perhaps it was a “boyish prank.” It turned out that Vosberg’s leg had a “preexisting condition” and experienced unexpectedly bad symptoms as a result of the kick. We use the case in law school as an early example of courts formally invoking the “eggshell-skull” rule. “You take your plaintiff as you find her.” After a mere three trips to the Wisconsin Supreme Court (grappling with, among other things, whether a physician should have been allowed to testify by deposition where facts already proven were excluded from the physician’s consideration in a hypothetical question), it was determined that—yes—Putney was liable in battery.
This got me thinking about the horseplay rules in workers’ compensation (because even though the kick was adjudged a “battery” the majority suggested it might have been deemed “horseplay” had it been delivered in the schoolyard, though it was muddy about how that would have changed the outcome). As many readers of this blog will know, horseplay in workers’ compensation is most frequently treated as a “course of employment” problem. That is, it is most often seen as a potential disqualifying “deviation” from employment that is assessed (as the Larson’s treatise explains) by considering:
(1) the extent and seriousness of the deviation, (2) the completeness of the deviation, (3) the extent to which the practice of horseplay had become an accepted part of the employment in question, and (4) the extent to which the nature of the employment may be expected to include some such horseplay. (In Wyoming, our lead case involves two minors scuffling in a Burger King).
As Larson’s also notes, in early workers’ compensation law non-participating victims of horseplay were uniformly denied coverage. Yet this denial is perfectly consistent with the traditional conception of workers’ compensation as a quid pro quo for negligence (not intentional tort) cases. As the treatise observes, “[j]ust as malicious assaults by co-employees were ruled out [of coverage] as intentional and personal, so sportive assaults were treated as something foreign to the inherent risks of the employment.” This is, of course, quite correct, but, more macroscopically, horseplay involves intentional conduct by somebody. Intentional conduct is not negligence (and it is obviously not accidental). It makes perfect sense that early workers’ compensation statutes, observing close fidelity to the precise quid pro quo scheme, would not have awarded compensation in what were essentially battery scenarios (at least seen from the perspective of the non-instigator). It also may explain in part why, even where horseplay has been determined an “incident of employment,” several jurisdictions continue to maintain the workers’ compensation rule that, only non-instigators of horseplay are covered. In a tort-battery regime, the instigator could have been the perpetrator of a battery. Awarding compensation to non-instigators on an “incident of employment” theory may, initially at least, have been a kind of compromise necessary to maintain the integrity of the system.
It may also be worth mentioning that the continued instigator/non-instigator split on horseplay in workers’ compensation corresponds interestingly with developments concerning the tort liability of employers for the intentional acts of their employees. The general rule historically has been that an employer is almost categorically not vicariously liable for the intentional tort of an employee who is not actuated by a purpose to serve the employer and motivated solely by a desire to satisfy the employee's own purposes. (For example, an employee's intentionally criminal conduct typically indicates a departure from conduct within the scope of employment). But under a more holistic “characteristic activities” standard several courts have concluded that some otherwise intentional employee conduct can be attributed to employers. “Sure, it looks like an intentional tort but this person does stuff like this all the time, and the employer knew it.” (Think of the bar room bouncer). This idea expands the potential for employer liability for intentional torts like batteries, and it makes sense that where employer liability for battery expands some courts/jurisdictions might respond by attempting to cover obviously “intentional conduct” within the workers’ compensation rubric (by, for example, requiring super-duper intentional conduct to remove an injury from the workers’ compensation lockbox). I would also note that Wyoming’s statute has always explicitly conferred immunity on employers for all tortious conduct, which makes for good sport for this workers’ compensation/torts prof: my employer comes into work one day and intentionally shoots me in the leg. Discuss.
It also makes sense that states with expanded battery liability would be more likely to include horseplay instigators as well as non-instigators into the workers’ compensation fold.
The moral of the story? Hypothetical questions are cool! No, that’s not it. I know. Be careful when you kick: you may set off an entire chain of legal history!
Back to course prep!
Michael C. Duff
Sunday, August 4, 2019
My work on a Wyoming workers’ compensation treatise (submitted to the publisher two weeks ago) has emphasized a fairly stark reality: while 90% of Wyoming workers are covered by workers’ compensation only 68% of that 90% work for employers required to provide workers’ compensation coverage. From the official Wyoming report tracking these numbers: “Ninety percent (90%) or 235,278 are covered by workers’ compensation. 175,455 employees are working in required coverage or extra-hazardous occupations, with 59,823 employees working in optional coverage positions.” Furthermore, “Currently, there are 22,629 employers in Wyoming . . . 16,322 or approximately 72% of those employers are registered with Workers’ Compensation (WC) as policyholders. Of the 16,322 employers who are registered, 11,027 or 68% have required coverage and 5,295 or 32% have optional coverage.”
So 68% of 90% yields a figure of 61% of employees working for employers that are required to carry workers’ compensation (monopolistic) coverage.
Like the situation in Texas, employers not covered by workers’ compensation are usually liable in tort. Also like in Texas, there are significant groupings of employees employed by companies utilizing alternative benefit plans (Walmart is the major example). Those plans are covered by ERISA and thus generally cannot be sued by employees who have elected to participate in them (and many employees will choose participation once they realize their employer is not covered by the state workers’ compensation system). Employees working for an employer that is not covered by workers’ compensation, and has not established an alternative benefit plan, find themselves in “Texas-land.” (Texas has operated an elective, opt-in system since the inception of workers’ compensation). The difficulty of pursuing tort claims is, of course, why we have workers’ compensation in the first place and, to add insult to injury, employees routinely find themselves required to arbitrate workplace injury tort disputes, even assuming they can find a theory to evade ERISA preemption.
Much of the coverage outcome in Wyoming has been driven by the narrowed statutory jurisdiction of the Wyoming statute, which applies, uniquely, only to “extrahazardous” employment (though Illinois also still technically possesses such a statute). Explicitly included extrahazardous “employments” are rather non-transparently defined with reference to the North American Industry Classification System. It takes a good deal of digging to come to the 61% coverage figure. But, as I have said in my treatise, any workers’ compensation system—however unusual—has little to recommend it if it excludes the largest employer in the state (Walmart). As the Larson’s treatise discusses, at 6 Larson's Workers' Compensation Law § 77.01, in the context of analyzing the Wyoming Supreme Court’s opinion in Araguz v. State, ex rel. Wyoming Workers’ Safety and Comp. Div., 2011 WY 148, 262 P.3d 1263 (Wyo. 2011):
Wal-Mart had been assigned [NAICS] code 452910, indicating “Warehouse Clubs and Supercenters.” Because such classification was not defined by Wyoming statute as “extrahazardous,” Wal-Mart [is] not required to participate in the state fund. Wal-Mart maintain[s] its own privately funded workers’ compensation fund, the Wal-Mart Plan, for the benefit of its employees who are injured on the job. The employees contend that the State’s failure to provide benefits was unconstitutional and improper as a matter of law. The court indicated that it would not address the appellants’ constitutional argument as the appeal was not an appropriate avenue for resolution of such an issue.
The entire structure is evocative of the path-breaking 2012 article by Professors Spieler and Burton, “The lack of correspondence between work‐related disability and receipt of workers' compensation benefits.” In Part II of the article, Spieler and Burton observed:
. . .[W]hile exclusions mean that the common law tort system is not supplanted by workers' compensation, exclusion also means that these workers can obtain benefits only through tort litigation (in which there must be proof of actual negligence), from private sources such as employment‐based health and disability plans, or from public programs, including SSDI, SSI, Medicaid, and Medicare. Some injured workers experience “dual denial”—they are eligible neither for damages from civil litigation nor benefits from alternative programs.
An elective workers’ compensation system like Wyoming’s may not expressly exclude (non-extrahazardous) employment from coverage; but Wyoming’s current economic downturn may ultimately render the distinction academic for a worker employed by an employer deciding to rescind its election to be covered by workers’ compensation. When profit margins become razor thin, an operator (now short term because of the onrush of events) may be willing to take a chance on a tort suit, especially if it figures out the arbitration gambit and the Supreme Court continues its FAA jurisprudence of arbitration uber alles.
Michael C. Duff