Humanistic Engineering: Engineering for the People

Everything is Too Complex

For some reason, engineers have decided that everything should be really complex. Complexity can be defined in a number of different ways, but it ultimately describes a condition that is difficult for humans to intellectually manage [11]. Complexity may sometimes be necessitated by the problem being solved or the application domain. However, often it is introduced as part of the design process due to overengineering, the trendiness of current engineering paradigms [e.g., using artificial intelligence (AI) and/or machine learning wherever possible], cost (“it’s cheaper to do it in software”), feature creep, or the convenience of “hacky” solutions.

The detrimental effect of complexity on system safety and stability has been known for decades [12], [13], [14]. Complexity is dangerous because it makes it difficult to identify bugs or design flaws, hard for any engineer to keep the entire system in his or her head, and hard for humans to use the system without making errors. As such, complexity makes every part of the system life cycle (evaluation, design, analysis, implementation, maintenance, and decommission) more difficult. Not surprisingly, complexity was a major factor in large-scale failures such as the Three Mile Island nuclear meltdown and the Bhopal India disaster (where a Union Carbide pesticide plant poisoned roughly 500,000 people due to a catastrophic gas leak). An active research area (which includes much of my own work) attempts to discover engineering solutions that can manage or mitigate the safety problems caused by complexity.

Safety is not the only problem that arises from systems being too complex. Complexity can encourage bad practices by hiding mistakes and incompetence: if nobody can figure out why something is not working, no individual is blamed. Complexity can hide bad intentions, that is, if the system is doing something that designers do not want others to see or notice, this can be hidden inside a complex process. This is a common practice of con artists as well as the financial industry. If you do not want regulators, oversight, competitors, or “marks” to figure out what is actually going on, hide the nefarious behavior in an extremely complex process. Some high-profile examples of this sort of practice can be seen in the iPhone tracking scandal from 2011, [64] where a software update resulted in people’s location being tracked and stored without their permission, as well as the Volkswagen emissions cheating scandal, [65] where the automaker hid a subroutine in onboard software that could detect laboratory emissions tests and alter car performance in response. Finally, complexity is both bad for democracy and the environment. Complexity is antidemocratic because it makes it difficult for people from different walks of life to understand how a system works, learn from it, repair it, or repurpose it [6], [15], [17]. This type of complexity is often deliberate because companies would rather have customers or clients purchase new products or authorized services rather than use self-repair or third-party services, both of which fail to earn profit for the original company. This is bad for the environment because it creates tons of (often high-tech) waste that could have been repaired or otherwise remanufactured.

Good Work is Too Reactive

The above discussion may give the impression that no good engineering work is being done. This is definitely not the case. There is considerable effort in trying to solve societal and technological problems. The issue with such undertakings is that they are reactive. Typically, the engineering effort identifies a specific problem and tries to fix it. Such a reactionary approach does not fundamentally address the underlying systemic issues that are the source of the problem. This means that reactive work too often only serves to prop up broken, poorly conceived, or unethical systems, frequently by adding additional layers of complexity.

For example, many have noticed that machine-learning algorithms, when used in applications like mortgage lending, make recommendations that we would consider sexist or racist. In response, researchers are scrambling to figure out how these biases can be removed from machine-learning methods or from input data [18]. But this reaction is largely missing the point. The purpose of using machine learning in this example is to identify the stereotype of the person applying for a mortgage and, in response, recommend a decision that best suits the profits of the organization. Thus, an algorithm that makes sexist or racist recommendations is actually doing what it is supposed to do. It may well be possible to find methods for removing particular biases from the machine-learning process. However, what is troubling about the original behavior is not just that it makes decisions that offend modern sensibilities, but that the machine-learning approach, by virtue of relying on stereotyping, is inherently inequitable and unjust.

This example also illustrates another problem with reactive work: it lacks vision. Research and developments that view solving problems as their motivation are constrained to a vision of the world that is only iteratively better than the current one, and one that is contextualized in terms of the original, broken system. So, returning to the machine-learning mortgage example, the problem-based reactive approach seeks to remove problematic bias from the algorithm. A more visionary and systemic perspective would attempt to figure out how everybody in society could obtain financing for purchasing a home or (even better) ensuring that everybody has adequate housing.

In situations where engineering work is more genuinely driven by a vision, it is usually done in a very narrow context. What seems to animate most engineers’ vision of the future is cool, shiny tech. In particular, we seem to be driven by a vision of making 1960s, 1970s, 1980s, and 1990s popular science fiction technology a reality. In this light, it makes sense that so much research is shifting toward AI, virtual reality, robotics, exoskeletons, self-driving cars, flying vehicles, and so on. While it is true that these things may help solve some problems, this is not really the point. The point is that the tech is “cool” and makes us feel like we are realizing a vision of the future from our childhoods. While this perspective will certainly have an impact on society, it is a very limited vision. It is one where the world effectively functions the way it does today, but with new toys and more convenience. In the absence of a vision of a better future, and technology specifically designed to realize that future, technology will only serve to perpetuate the status quo, which includes its regressive tendencies.

We Do Not Design Things For People

The overemphasis on tech innovation is creating a paradoxical situation where an immense amount of engineering is not designing things for people. Rather, products are created to fuel the technological and monetary ambitions of our larger institutions. This is why so much modern work infrastructure (online reimbursement systems, course administration software, electronic medical records) is so difficult to use: it was not designed to help people do their jobs.

In many cases, human concerns, rather than being the point of engineering, are viewed as an impediment to its progress. Autonomy engineers are mounting significant efforts to determine how to make people defer to and “trust” automation. Control theorists view humans as a source of error and instability. Cybersecurity engineers view end users as the “enemy” of security, the equivalent of hackers causing problems. Modern tech companies view humans as sources of data and money and thus any sort of effective human-centered design (which is almost exclusively focused on the consumer space) is meant to extract resources from them as easily (and often as covertly) as possible. But even in this space, devices like mobile phones are supposed to facilitate media consumption and textual communication but make us interact on tiny, smudgy screens with awkward touchscreen keyboards.

Complex technologies that attempt to automate human work have a number of problems that have been recognized for decades [19], [20], [21]: humans often have their roles changed to ones (such as monitoring the automation) that are incompatible with human cognition1; automation can be brittle and thus fail in situations unanticipated during design; the human may not be able to track the state of the system, leading to poor situation awareness, mode confusion, disorienting automation surprise, and human errors; humans can become too reliant on automation and use it inappropriately; and humans who cannot determine why the automation is doing what it is doing may distrust it and not use it when it is appropriate. Despite the well-established nature of these problems, they have persisted into modern times [23], [24]. Given the complex and unexplainable nature of emerging autonomous and machine-learning technologies, these issues are expected to get even worse than they were in previous generations [25].

Engineering Lacks Values

Finally, the core issue that, I suspect, drives all of the other problems is that engineering lacks values. In my experience, most engineers (and researchers within the field) view engineering knowledge and technology as goods in their own right and that by bringing these into the world, society and the human condition are being advanced.

It does not take much historical searching to find good evidence that science and engineering are not inherently virtuous. The Nazis were able to execute the holocaust as effectively as they did through the use of (then) cutting-edge punched card technology [26]. They were also clearly devoted to advancing the science and technology of industrial operations, nuclear physics, and rocketry to support their authoritarianism and war efforts. The problems outlined at the beginning of this article also make it clear that this is not just a historical anomaly. Thus, it is more accurate to see science and technology as tools that, unless special effort is taken to the contrary, will be used to support and advance the goals of society’s powerful.

The net effect is that engineering, often despite the intentions of its practitioners, is contributing to a society that is less fair, less just, less democratic, disempowering, and more precarious.

So, if engineering advances are primarily directed by the ideology of society’s powerful, then neoliberalism [27] is our driving ideology.

Core Principles of Humanistic Engineering

The specific goals of humanistic engineering projects will vary and evolve over time. I have identified five core principles that, together, should provide a comprehensive value system for ensuring that these goals are pursued with a common set of humanistic values.

Realizing Humanistic Engineering

On its face, it seems ridiculous that humanistic engineering is not the way that engineering is currently done. But as it is, humanistic engineering would constitute a major transformation. It is my intent for humanistic engineering to be the dominant engineering paradigm. Achieving this will require a significant multidisciplinary effort with advances in research, education, and service dimensions.

Improved Creativity

Despite the current emphasis on STEM education and research, we have been seeing significant declines in the rate of scientific innovation [32], [33], with increased complexity being cited as a critical factor in the slowdown [32]. Interestingly, a loss of creativity is a common, broad criticism of neoliberalism [34]. This is because the inherent complexity of everything, the pervasiveness of seemly intractable institutions, and the established market success of previous products limits people’s ability to imagine alternatives.

Thus, humanistic engineering, by stepping outside of the neoliberalism perspective, has the potential to spur creativity. By focusing on systemic problems and humanistic goals, rather than technological or reactive incrementalism, new solutions and unique technologies will be developed. Innovation and creativity will also be spurred by the fact that humanistic engineering will inherently require the incorporation of new concepts from the liberal arts and humanities.

Increased Entrepreneurship

Neoliberal proponents espouse the importance of entrepreneurship in driving innovation, improving people’s material conditions, and growing the economy. However, under neoliberalism, entrepreneurship has been declining for more than a decade, even in the high-tech sector [35]. This seeming paradox actually makes sense in light of the complex, techcentric, technocratic, and market-based solutions favored by neoliberalism. By making products unnecessarily complex and technological, companies make it harder to develop secondary entrepreneurial markets for service, repair, modification, or remanufacture. If the original company has significant influence, they may encourage laws or adopt policies that make it illegal or a violation of warranty to do these things. Often, companies will create unique tools to perform repairs that they only make available to authorized parties or people paying them for service agreements. Finally, due to economic deregulation, any large company that perceives a new entrepreneurial effort as a threat can use its superior resources and influence within the marketplace to drive the newcomer out of business. For example, a company could refuse to do business with a customer that uses or stocks a competitor’s product.

By encouraging scientific creativity, we would expect humanistic engineering to push entrepreneurship in new directions. By enabling democracy in engineered products, we also expect it to encourage a host of third-party industries to grow around major innovations. Thus, humanistic engineering should stimulate entrepreneurship and facilitate its associated benefits.

Diversity

While not explicitly represented in the core principles, humanistic engineering should have a positive impact on diversity in engineering. There are several reasons for this. First, diversity and improved social outcomes are compatible with humanistic engineering’s second and fourth core principles from above. Second, in STEM areas that engage with human issues more directly (like medicine), gender disparities have been closed much more effectively than they have been in engineering [36]. Thus, I suspect that a major part of the problem engineering has with diversity is a direct result of its lack of humanism. Beyond this, underrepresented minority populations have disproportionately been on the sharp end of all of the damaging trends highlighted in the introduction. Given the role that technology has played in these outcomes, it would make sense that these groups would not inherently see engineering as an uplifting or positive force in their communities. Furthermore, there have been studies that have shown that a higher percentage of minorities (as compared to nonminorities) pursue STEM for the purpose of affecting social change [37] and that the “narrow focus on decontextualized science” [38] often discourages them once they have enrolled [38], [39]. This suggests that aligning engineering with humanistic pursuits will help recruit and retain unrepresented populations. Third, humanistic engineering specifically encourages worker and population empowerment, especially as it relates to unionization and collective action. Evidence shows that labor unions have positive effects on diversity and reduce discrimination: union membership decreases racial resentment among white workers [40]; minorities often join unions explicitly for protection from discrimination [41]; and union membership generally reduces racial and gender wage gaps [41], [42], [43]. This actually makes a fair amount of sense. Unions are based on the idea that people with different backgrounds and perspectives come together in solidarity based on shared goals and interests. This ultimately helps break down barriers by giving common ground to people that may otherwise have trouble relating to each other.

In this light, humanistic engineering should both facilitate diversity within engineering and help correct discriminatory practices and outcomes across society.

HCI and Human Factors Engineering

HCI, and human factors engineering have all made, to some extent, contributions to human-centered engineering and accessibility. HCI is, by its nature, primarily concerned with humans interacting with computers. As such, HCI largely serves the advancement and adoption of computation, not humanistic goals. Thus, when human goals and computational advancement are aligned, HCI can indeed be a humanistic undertaking. However, as we have already discussed, the technology focus generally makes such alignment difficult. Human factors engineering (my primary research discipline) is probably the closest conceptually to humanistic engineering, but with a very narrow scope. Human factors engineering is principally focused on engineering safe and effective human work. This means that human factors emphasizes the engineering of things in a human-centered way in so much that it helps people do their jobs. This inherently limits the scope of the goals that are considered and often aligns human factors engineers with management: human-centered engineering is worthwhile if it contributes to a more profitable workplace [44].3 Importantly, both areas are marginal subdisciplines whose influence on most engineering projects only comes near the end of the design phase, when the human-centered perspective will have minimal impact on the overall project. Humanistic engineering makes human-centeredness central to all engineering. This is fundamentally different because it increases the scope of human-centered practice and, thus, will help ensure that a broader set of human goals are accounted for at all steps in the engineering process.

Humanitarian Engineering

Humanitarian engineering [46] is a multidisciplinary application area of engineering focused on improving the condition of disadvantaged populations, particularly in response to disasters in the developing world. Humanitarian engineering falls into what I previously classified as reactive research: genuinely noble pursuits that iteratively address problems without fixing systemic causes. Additionally, as an application area, humanitarian engineering is not a comprehensive reformulation of engineering the way humanistic engineering is. Humanitarian efforts will definitely be part of humanistic engineering. However, these will be conducted as part of a larger framework that attempts to correct the deep systematic problems that produced the disparities. It will also do so while avoiding the destructive implications of economic imperialism that can accompany humanitarianism.

Engineering Ethics and Engineering Justice

Ethics has long been a part of the engineering canon. It has been a requirement of ABET accreditation since 1997 [47] and organizations like the National Society of Professional Engineers (NSPE) have adopted codes of ethics [69]. Engineering justice [48] is an education initiative that seeks to connect engineering to social justice. Both are synergistic with humanistic engineering in that all seek to broaden the perspective of engineers to include societal concerns. However, humanistic engineering is different. It expands the scope of all engineering disciplines to encompass its core principles. Furthermore, humanistic engineering is a fundamental reformulation of the discipline, not an “add-on” like both engineering ethics and justice. Additionally, despite its legacy, engineering ethics has clearly been insufficient to address the problems discussed in the introduction. There are several potential reasons for this. First, even within the established ethical codes, there are possible conflicts. Within the NSPE code, engineers are expected to both “hold paramount the safety, health, and welfare of the public” and “act for each employer or client as faithful agents or trustees.” This can result in a dilemma where an engineer must choose between professional obligation and the welfare of the public. Second, engineering ethics does something that humanistic engineering explicitly seeks to avoid: putting the responsibility of organizational ethical behavior on the individual. That is, if an engineer encounters an ethical dilemma with a project, it is his or her personal responsibility to do something about it. In the NSPE code, an engineer is instructed to report up the chain of authority (the client organization, governing bodies, law enforcement, etc.) if the organization is doing something unethical. This puts the individual engineer in a very precarious situation where he/she/they can choose to either keep quiet or potentially face retribution for speaking out. In fact, there is good evidence that this relationship between engineers and employers suppresses ethical behavior in practice.

A survey of UC San Diego undergraduates found that power differentials between employees and employers were a major concern and compromise for socially conscious engineers on the job market [49]. Similarly, a survey of University of Colorado graduates found that engineers who left the profession were more likely to say they wanted a job with an “ability to contribute to society” [50]. Because engineering justice is based on the same education strategy as ethics, we expect it to have similar shortcomings. Humanistic engineering addresses the deficiencies of these areas by adopting a democratic model of power and training engineers to wield it for ethical and humanistic ends. Furthermore, by democratically empowering people with technology, humanistic engineering gives the population the power to drive society in ethical directions, taking the further burden off of individual engineers and citizens.

Design

Within the design discipline, there have been a number of advances that include human-, environmental-, and value-focused principles in products. For example, user-centered and human-centered design attempt to account for the end user and human goals [51]. Recent developments extend these concepts beyond humans to account for the environment and value frameworks [52], [53], [54]. Concepts related to worker power have also played a key role in computer-supported cooperative work and participatory design, both concerning themselves with determining how to have people work democratically towards shared goals [55], [56]. The latter specifically focuses on the importance of end-user participation in the design process and has strong roots in the Scandinavian labor movement. Universal design [57] seeks to create artifacts that are accessible and inclusive to as many people as possible. Finally, value-sensitive design [58] is an approach that incorporates stakeholder values into the design process.

All of these developments have clear roles to play in humanistic engineering. However, there are important limitations. In particular, design focuses almost exclusively on what the technology will be like, not how it will be realized. Both the what and the how are accounted for in engineering. As previously stated, how a design is made to work will be critical to the democratization and empowerment capabilities of technology. Thus, humanistic engineering adds critical considerations. Furthermore, many cooperative, democratic, and inclusive design principles are rooted in stakeholder analysis. This is meant to ensure that different perspectives are represented. This is also common practice in business, where the market-driven values of neoliberalism dominate. A good example of how this can manifest as a limitation is seen in value-sensitive design, where no specific commitment to a set of values is made [59]. As such, making money or dominating a market can be treated as values along with (or instead of) other, more humanistic principles. Thus, by attempting to specifically combat neoliberal influence in technology development, humanistic engineering makes a novel contribution.

Problems Outlined are Bigger than Engineering

One argument I have encountered when discussing the ideas in this article with colleagues is that the problems I have identified are bigger than engineering. This particular argument, is, I fear, rooted in a desire for engineers to avoid responsibility for the fact that technology has not accomplished the benefits it has promised. Engineering is the discipline that translates scientific developments into society. If science and technology are the instruments of human progress that engineers claim they are, then engineering is particularly well suited to take the lead in ensuring that the promised benefits are actually realized. That said, humanistic engineering is, prima facie, a multidisciplinary undertaking and any other disciplines that want to contribute are welcome.

Values/Politics Should Stay Out of Science and Engineering

The other concern colleagues have raised with me is that some individuals may not share humanistic values. Also, given the political nature of the labor power emphasis of humanistic engineering, I expect objections on the grounds that science should avoid politicization. There are two problems with these arguments. First, engineering already claims that it serves to advance the human condition. Because this ambition is not actually a fundamental part of the engineering process, reality produces divergent results. In this context, humanistic engineering is attempting to clarify the values that engineering should be pursuing (a distillation of what is already claimed), while giving it the actual ability to adhere to them. Second, I hope it is now clear, that in the absence of engineers asserting their values through their work, their work will serve the values of other powerful societal interests. Thus, there is no neutral position. Engineering work will always support ideological and political perspectives. Humanistic engineering offers us the ability to take agency over the values asserted by our work so that it can be used for real good.

All of this is not to say that humanistic engineering supports indoctrination. Should any engineers choose to forgo upholding the standards of humanistic engineering, there will be no enforcement that strips them of credentials, blacklists them from the profession, or otherwise “cancels” them. Such practices are authoritarian and run contrary to humanistic engineering’s perspectives on democracy and empowerment. There are, and likely always will be, enormous opportunities for those who choose to forgo civic responsibility to support the interests of the moneyed and powerful. Humanistic engineering will simply give those of us who wish to take a different course more opportunities to do so.