Stealth Hiring Quotas Via AI

24 Monday Oct 2022

Posted by Nuetzel in Discrimination, Diversity, Quotas, Uncategorized

Tags

AI, AI Bill of Rights, Algorithmic Bias, Algorithms, American Data Privacy and Protection Act, Artificial Intelligence, DEI, Disparate impact, Diversity Equity Inclusion, EEOC, Hiring Quotas, Machine Learning, Neural Networks, Protected Classes, Stealth Quotas, Stewart Baker, Volokh Conspiracy

Hiring quotas are of questionable legal status, but for several years, some large companies have been adopting quota-like “targets” under the banner of Diversity, Equity and Inclusion (DEI) initiatives. Many of these so-called targets apply to the placement of minority candidates into “leadership positions”, and some targets may apply more broadly. Explicit quotas have long been viewed negatively by the public. Quotas have also been proscribed under most circumstances by the Supreme Court, and the EEOC’s Compliance Manual still includes rigid limits on when the setting of minority hiring “goals” is permissible.

Yet large employers seem to prefer the legal risks posed by aggressive DEI policies to the risk of lawsuits by minority interests, unrest among minority employees and “woke” activists, and “disparate impact” inquiries by the EEOC. Now, as Stewart Baker writes in a post over at the Volokh Conspiracy, employers have a new way of improving — or even eliminating — the tradeoff they face between these risks: “stealth quotas” delivered via artificial intelligence (AI) decisioning tools.

Skynet Smiles

A few years ago I discussed the extensive use of algorithms to guide a range of decisions in “Behold Our Algorithmic Overlords“. There, I wrote:

“Imagine a world in which all the information you see is selected by algorithm. In addition, your success in the labor market is determined by algorithm. Your college admission and financial aid decisions are determined by algorithm. Credit applications are decisioned by algorithm. The prioritization you are assigned for various health care treatments is determined by algorithm. The list could go on and on, but many of these ‘use-cases’ are already happening to one extent or another.”

That post dealt primarily with the use of algorithms by large tech companies to suppress information and censor certain viewpoints, a danger still of great concern. However, the use of AI to impose de facto quotas in hiring is a phenomenon that will unequivocally reduce the efficiency of the labor market. But exactly how does this mechanism work to the satisfaction of employers?

Machine Learning

As Baker explains, AI algorithms are “trained” to find optimal solutions to problems via machine learning techniques, such as neural networks, applied to large data sets. These techniques are are not as straightforward as more traditional modeling approaches such as linear regression, which more readily lend themselves to intuitive interpretation of model results. Baker uses the example of lung x-rays showing varying degrees of abnormalities, which range from the appearance of obvious masses in the lungs to apparently clear lungs. Machine learning algorithms sometimes accurately predict the development of lung cancer in individuals based on clues that are completely non-obvious to expert evaluators. This, I believe, is a great application of the technology. It’s too bad that the intuition behind many such algorithmic decisions are often impossible to discern. And the application of AI decisioning to social problems is troubling, not least because it necessarily reduces the richness of individual qualities to a set of data points, and in many cases, defines individuals based on group membership.

When it comes to hiring decisions, an AI algorithm can be trained to select the “best” candidate for a position based on all encodable information available to the employer, but the selection might not align with a hiring manager’s expectations, and it might be impossible to explain the reasons for the choice to the manager. Still, giving the AI algorithm the benefit of the doubt, it would tend to make optimal candidate selections across reasonably large sets of similar, open positions.

Algorithmic Bias

A major issue with respect to these algorithms has been called “algorithmic bias”. Here, I limit the discussion to hiring decisions. Ironically, “bias” in this context is a rather slanted description, but what’s meant is that the algorithms tend to select fewer candidates from “protected classes” than their proportionate shares of the general population. This is more along the lines of so-called “disparate impact”, as opposed to “bias” in the statistical sense. Baker discusses the attacks this has provoked against algorithmic decision techniques. In fact, a privacy bill is pending before Congress containing provisions to address “AI bias” called the American Data Privacy and Protection Act (ADPPA). Baker is highly skeptical of claims regarding AI bias both because he believes they have little substance and because “bias” probably means that AIs sometimes make decisions that don’t please DEI activists. Baker elaborates on these developments:

“The ADPPA was embraced almost unanimously by Republicans as well as Democrats on the House energy and commerce committee; it has stalled a bit, but still stands the best chance of enactment of any privacy bill in a decade (its supporters hope to push it through in a lame-duck session). The second is part of the AI Bill of Rights released last week by the Biden White House.”

What the hell are the Republicans thinking? Whether or not it becomes a matter of law, misplaced concern about AI bias can be addressed in a practical sense by introducing the “right” constraints to the algorithm, such as a set of aggregate targets for hiring across pools of minority and non-minority job candidates. Then, the algorithm still optimizes, but the constraints impinge on the selections. The results are still “optimal”, but in a more restricted sense.

Stealth Quotas

As Baker says, these constrains on algorithmic tools would constitute a way of imposing quotas on hiring that employers won’t really have to explain to anyone. That’s because: 1) the decisioning rationale is so obtuse that it can’t readily be explained; and 2) the decisions are perceived as “fair” in the aggregate due to the absence of disparate impacts. As to #1, however, the vendors who create hiring algorithms, and specific details regarding algorithm development, might well be subject to regulatory scrutiny. In the end, the chief concern of these regulators is the absence of disparate impacts, which is cinched by #2.

About a month ago I posted about the EEOC’s outrageous and illegal enforcement of disparate impact liability. Should I welcome AI interventions because they’ll probably limit the number of enforcement actions against employers by the EEOC? After all, there is great benefit in avoiding as much of the rigamarole of regulatory challenges as possible. Nonetheless, as a constraint on hiring, quotas necessarily reduce productivity. By adopting quotas, either explicitly or via AI, the employer foregoes the opportunity to select the best candidate from the full population for a certain share of open positions, and instead limits the pool to narrow demographics.

Demographics are dynamic, and therefore stealth quotas must be dynamic to continue to meet the demands of zero disparate impact. But what happens as an increasing share of the population is of mixed race? Do all mixed race individuals receive protected status indefinitely, gaining preferences via algorithm? Does one’s protected status depend solely upon self-identification of racial, ethnic, or gender identity?

For that matter, do Asians receive hiring preferences? Sometimes they are excluded from so-called protected status because, as a minority, they have been “too successful”. Then, for example, there are issues such as the classification of Hispanics of European origin, who are likely to help fill quotas that are really intended for Hispanics of non-European descent.

Because self-identity has become so critical, quotas present massive opportunities for fraud. Furthermore, quotas often put minority candidates into positions at which they are less likely to be successful, with damaging long-term consequences to both the employer and the minority candidate. And of course there should remain deep concern about the way quotas violate the constitutional guarantee of equal protection to many job applicants.

The acceptance of AI hiring algorithms in the business community is likely to depend on the nature of the positions to be filled, especially when they require highly technical skills and/or the pool of candidates is limited. Of course, there can be tensions between hiring managers and human resources staff over issues like screening job candidates, but HR organizations are typically charged with spearheading DEI initiatives. They will be only too eager to adopt algorithmic selection and stealth quotas for many positions and will probably succeed, whether hiring departments like it or not.

The Death of Merit

Unfortunately, quotas are socially counter-productive, and they are not a good way around the dilemma posed by the EEOC’s aggressive enforcement of disparate impact liability. The latter can only be solved only when Congress acts to more precisely define the bounds of illegal discrimination in hiring. Meanwhile, stealth quotas cede control over important business decisions to external vendors selling algorithms that are often unfathomable. Quotas discard judgements as to relevant skills in favor of awarding jobs based on essentially superficial characteristics. This creates an unnecessary burden on producers, even if it goes unrecognized by those very firms and is self-inflicted. Even worse, once these algorithms and stealth quotas are in place, they are likely to become heavily regulated and manipulated in order to achieve political goals.

Baker sums up a most fundamental objection to quotas thusly:

“Most Americans recognize that there are large demographic disparities in our society, and they are willing to believe that discrimination has played a role in causing the differences. But addressing disparities with group remedies like quotas runs counter to a deep-seated belief that people are, and should be, judged as individuals. Put another way, given a choice between fairness to individuals and fairness on a group basis, Americans choose individual fairness. They condemn racism precisely for its refusal to treat people as individuals, and they resist remedies grounded in race or gender for the same reason.”

Quotas, and stealth quotas, substitute overt discrimination against individuals in non-protected classes, and sometimes against individuals in protected classes as well, for the imagined sin of a disparate impact that might occur when the best candidate is hired for a job. AI algorithms with protection against “algorithmic bias” don’t satisfy this objection. In fact, the lack of accountability inherent in this kind of hiring solution makes it far worse than the status quo.

Robots and Tradeoffs

31 Wednesday Aug 2016

Posted by Nuetzel in Living Wage, Markets, Technology

≈ 1 Comment

Tags

AI, Artificial Intelligence, Biohacking, Capital Intensity, Collaborative Automation, Cyborgization, David Autor, Don Boudreaux, Exoskeletons, Factor Substitution, Farmbots, George Selgin, gier.org, Hal Varian, Human Augmentation, Labor Productivity, Minimum Wage, Owning Machines, Pew Research Center, Resource Allocation, Robert Samuelson, Robotics, Scarcity, Singularity Hub, Superabundance, Work-Leisure Tradeoff

The other day, a few colleagues were lamenting the incipient robot domination of the workplace. It is true that advances in automation and robotics are likely to displace workers in a variety of fields over the next few decades. However, the substitution of capital for labor is not a new phenomenon. It’s been happening since the start of the industrial age. At the same time, capital has been augmenting labor, making it more productive and freeing it up for higher-valued uses, many of which were previously unimagined. The large-scale addition of capital to the production process has succeeded in raising labor productivity dramatically, and labor income has soared as a consequence. That is likely to continue as increasingly sophisticated robots assume certain tasks entirely and collaborate with workers on others, even in the service sector.

Advanced forms of automation are another step in the progression of technology. The process itself, however, and the adoption of robotics, might well be hastened by public policy that pushes labor costs to levels not commensurate with productivity. I wrote about this process in “Automate No Job Before Its Time” on Sacred Cow Chips late last year. The point of that essay was that government-imposed wage floors create an incentive for automation. Because a wage floor has its impact at the bottom of the wage scale, at which workers are the least-skilled, this form of government intervention creates a regrettable and unnatural acceleration in the automation process. Other mandated benefits and workplace regulations can have similar effects.

Robert Samuelson makes the same points in “Our Robot Panic Is Overblown“. He notes the effectiveness of the U.S. economy in creating jobs over time in the presence of increasing capital-intensity. But he also warns of potential missteps, including the dangers of government activism:

“There are two dangers for the future. One is that the new jobs created by new technologies will require knowledge and skills that are in short supply, leaving unskilled workers without income and the economy with skill scarcities. … The second danger is that government will damage or destroy the job creation process. We live in a profit-making economic system. Government’s main role is to maintain the conditions that make hiring profitable. … If we make it too costly for private firms to hire (through high minimum wages, mandated costs and expensive regulations) — or too difficult to fire — guess what? They won’t hire. That’s what ought to worry us, not the specter of more robots.“

Historically, automation has actually created more jobs than it has destroyed. In general, however, the new jobs have required higher levels of skills than the jobs lost. In “Why Are There Still So Many Jobs? The History and Future of Workplace Automation“, David Autor of MIT says it this way:

“Automation does indeed substitute for labor—as it is typically intended to do. However, automation also complements labor, raises output in ways that lead to higher demand for labor, and interacts with adjustments in labor supply. … journalists and even expert commentators tend to overstate the extent of machine substitution for human labor and ignore the strong complementarities between automation and labor that increase productivity, raise earnings, and augment demand for labor.“

As with almost all automation, robots will replace workers in the most routine tasks. Tasks involving less routine will not be as readily assumed by robots. To a large degree, people misunderstand the nature of automation in the workplace. The introduction of robots often requires collaboration with humans, but again, these humans must have more highly-developed skills than a typical line worker.

Hal Varian, who is the chief economist at Google, describes the positive implications of the ongoing trend to automate (see the link in the last paragraph of this post), namely, less drudgery and more leisure:

“If ‘displace more jobs’ means ‘eliminate dull, repetitive, and unpleasant work,’ the answer would be yes. How unhappy are you that your dishwasher has replaced washing dishes by hand, or your vacuum cleaner has replaced hand cleaning? My guess is this ‘job displacement’ has been very welcome, as will the ‘job displacement’ that will occur over the next 10 years. The work week has fallen from 70 hours a week to about 37 hours now, and I expect that it will continue to fall. This is a good thing. Everyone wants more jobs and less work. Robots of various forms will result in less work, but the conventional work week will decrease, so there will be the same number of jobs (adjusted for demographics, of course).“

An extreme version of the “robot domination” narrative is that one day in the not-too-distant future, human labor will be obsolete. Automation is not limited to repetitive or menial tasks by any means. A wide variety of jobs requiring advanced skills have the potential to be automated. Already, robots are performing certain tasks formerly done only by the likes of attorneys, surgeons, and computer programmers. Robots have the potential to repair each other, to self-replicate, to solve high-order analytical problems and to engage in self-improvement. With advances in artificial intelligence (AI), might humans one day become wholly obsolete for productive tasks? What does that portend for the future of so many human beings and their dependents who, heretofore, have relied only on their labor to earn a living?

There are any number of paths along which the evolution of technology, and its relationship to workers and consumers, might play out. The following paragraphs examine some of the details:

The Human Touch: There will probably always be consumers who prefer to transact with humans, as opposed machines. This might be limited to a subsegment of the population, and it might be limited to the manufacture of certain artisan goods, such as hand-rolled cigars, or certain services. Some of these services might require qualities that are more uniquely human, such as empathy, and the knowledge that one is dealing with a human would be paramount. This niche market might be willing to pay premium prices, much as consumers of organic foods are willing to pay an extra margin. However, it will be necessary to retain the perceived quality of the human touch and to remain reasonably competitive with automated alternatives on price.

Human Augmentation: Another path for the development of technology is the cyborgization of labor. This might seem rather distasteful to current sensibilities, but it’s a change that is probably inevitable. At least some will choose it. Here is an interesting definition offered by geir.org:

“Cyborgization is the enhancement of a biological being with mechanical or non-genetically delivered biological devices or capabilities. It includes organ or limb replacements, internal electronics, advanced nanomachines, and enhanced or additional capabilities, limbs, or senses.“

These types of modifications can make “enhanced” humans competitive with machines in all kinds of tasks. The development of these kinds of technologies is taking place within the context of rehabilitative medical care and even military technology, such as powerful exoskeletons, but the advances will make their way into normal civilian life. There is also development activity taking place among extreme hobbyists underground, such as “biohackers” who perform self-experimentation, embedding magnets or electronic chips in their bodies in attempts to develop a “sixth sense” or enhanced physical abilities. Even these informal efforts, while potentially risky to the biohackers themselves, might lead to changes that will benefit mankind, much like the many great garage tinkerers who have been important to innovation in the past.

Owning Machines: Ownership of capital will take on a greater role in providing for lifetime earnings. Can the distribution of capital ever be broadened to the extent needed to replace lost labor income? There are ways in which this can occur. The first thing to note is that the transition to a labor-free economy, were that to transpire, would play out over many years. Second, we have witnessed an impressive diffusion of advanced technologies in recent decades. Today, consumers across the income spectrum hold computers in their pockets that are more than the equivalent of the supercomputers that existed 50 years ago. Today’s little computers are far more useful in many ways, given wireless internet connectivity. There are many individuals for whom these devices are integral to earning an income. Thus, the rate of return on these machines can be quite impressive. The same is true of computers, software (sometimes viewed as capital) and printers, not to mention other “modern” contrivances with income-earning potential such as cars, trucks and a vast array of other tools and hardware. Machines with productive potential will continue to make their way into our lives, both as consumers and as individual producers. This also will include value-added production of goods at home, even for use or consumption within the household (think 3D printers, or backyard “farmbots”).

Saving Constructs: Most of the examples above involve machines that require some degree of human collaboration. Of course, even the act of consuming involves labor: I must lift the fork to my mouth for every bite! But in terms of earning income from machines, are there other ways in which ownership of capital can be broadened? The first answer is an old one: saving! But there is no way most individuals at the start of their “careers” can garner a significant share of income from capital. Other social arrangements are probably necessary. One of great importance is the family and family continuity. Many who have contemplated a zero-labor future imagine a world in which there are only two kinds of actors: individuals and the state. Stable families, however, hold the potential for accumulating capital over time to provide a flow of income for their members. Other forms of social organization can fill this role, but they must be able to accumulate capital endowments across generations. Of course, in an imagined world with minimal opportunities for labor, some have concluded that society must collectively provide a guaranteed income. To indulge that view for just a moment, a world of complete automation would almost certainly be a world of superabundance, so goods would be extremely cheap. That means a safety net could be provided at a very low cost. Nonetheless, it would be far preferable to do so by distributing a minimal number of shares of ownership in machines. These shares would have some value, and to improve resource allocation, it should be the individual’s responsibility to manage those shares.

Economic Transition: The dynamics of the transition to robot-dominated production raise some interesting economic questions. Should the advancement of robotics and artificial intelligence create a massive substitution away from labor, it will be spurred by 1) massive upward shifts in the productivity of capital relative to its cost; and 2) real wages that exceed labor’s marginal productivity. There will be stages of surging demand for the kinds of advanced labor skills that are complementary to robots. The demand for less advanced labor services does not have to fall to zero, however. There will be new opportunities that cannot be predicted today. Bidding for scarce capital resources and the flow of available saving will drive up capital costs, slowing the transition. And as long as materials, energy and replacement parts have a cost, and as long as savers demand a positive real return, there will be a margin along which it will be profitable to employ various forms of labor. But downward adjustment of real wages will be required. Government wage-floor policies must be abandoned. That will not be as difficult as it might sound: the kind of automation envisioned here would have profound effects on overall costs and the supply of goods, leading to deflation in the prices of consumables. As long as real factor prices can adjust, there will almost certainly remain a balance between the amounts of capital and labor employed.

In “Robots Are Nothing New“, Don Boudreaux passes along this comment from George Selgin:

“I’ve always been aghast at finding many otherwise intelligent economists arguing as if technology had a mind of its own, developing willy-nilly, or even perversely, in relation to the relative scarcity of available factors, including labor. Only thus can it happen that labor-saving technology develops to a point where labor, instead of being relatively scarce, becomes superabundant!

The fundamental problem, I believe, is confusion of the role of technological change with that of government interference with the pricing of labor services that is among the things to which technology in turn responds. Labor-saving technology becomes associated with unemployment, not because the last is a consequence of the former, but because both are contemporaneous consequences of a common cause, to wit: minimum wage laws and other such interference that sets wage rates above their market-clearing levels.“

There is much disagreement on the implications of automation. This excellent survey of experts by the Pew Research Center contains a number of insights. Also, visit Singularity Hub for a number of great articles on automation and AI, some of which are surprising. I believe that these technologies hold a great deal of promise for humanity. The process will not take place as suddenly as some fear, but ill-conceived policies such as a mandated “living wage” would put us on an unnaturally speedy trajectory. Opportunities for the least-skilled workers will be foreclosed too soon, before those individuals can develop skills and improve their odds of establishing a life free of dependency. Too rapid an adoption of advanced automation and AI would increase the likelihood of choosing suboptimal production methods that might be difficult to change later, and it would leave little time for education and training for workers who might otherwise leverage new technologies. The benefits of automation and their diffusion can be maximized by allowing advances to take a natural course, guided by market forces, with as little interference from government as possible.