The Ethics and Macroeconomics of Automation
Finding purpose, work, and flow with our machine friends
For many people work is a source of purpose. This is because work itself can be purposeful and so too can it finance purposeful activities outside of work. Note that “purpose” means different things to different people: in more well-off areas like the US “purpose” can be partaking in more creative activities, in more developing areas “purpose” can be having better access to education. Regardless, purposeful activities make us more fulfilled in life which should be the end goal of what we do—a notion which Aristotle called the “final good”.1
Per this framework, advances in workplace tech should aim to (1) advance how purposeful our lives are within work or (2) advance how purposeful our lives are outside work. This relation between purpose, work, and technology has largely been advantageous in the past, where better tech has led to more purpose. For example, it is easier today to access computing resources and information than ever before, allowing for more analytical work. This has been the case because new tech has historically led to new jobs. However, this dynamic will fundamentally change if automation reaches the point where fewer jobs are created than are consumed. In light of this, companies which benefit most from automation are ethically obligated to invest in reskilling and reorienting displaced workers. Incentives must be aligned such that companies do not replace workers with machines strictly on an economics basis but rather focus on the longer-term, purpose-altering effects of automation. What does it mean if it is more efficient for a subset of the population to not work? How can we promote more equality in this scenario? Investing in education and reskilling today is a productive way to attack these problems.
1. Historically, automation has net-created jobs; this will likely change going forwards
Studies conducted in the past confirm that more jobs are created than are removed as tech advances.2 Companies like Amazon, the leaders in warehouse automation, are quick to provide their own evidence that they have added more jobs than they have automated away.3 This argument gives a skewed picture of reality: for a company growing net sales exponentially as it expands to new markets, adding jobs is simply scaling to meet demand.4 Additionally, this does not mean that Amazon is net adding jobs when considering companies which they outcompete along the way. The same can be said of Walmart, Ocado, and other automation-focused supply chain companies. These companies thus have a duty to hold themselves accountable to whatever societal shifts they cause on the consumer, technical, and labor fronts.
2. Manufacturing and supply chain companies determine the wellbeing of a significant portion of humanity
Regardless of the direction and magnitude of automation going forwards, the potential for impact is immense. Amazon employs more than 647,000 people.5 Walmart hires an astounding 2.2 million people.6 Tens of millions more work full-time as suppliers, contractors, and aggregators on top of the Amazon and Walmart marketplaces. Any shift to employment will have profound economic and societal implications on all direct and indirect employees. Given the scale of this impact, these companies are obligated to ensure stable, healthy navigation of these shifts. As it is, many employed by these companies are dissatisfied. Indeed, Amazon warehouse workers have described the work as grueling, depressing, and robotlike.7 For at least a subset of these people, their current professional state is suboptimal and could surely be improved with shifts in the labor market.
3. Assuming no technical risk, it makes economic sense for automation to net remove jobs in the long-run
Technology has, quite obviously, significantly shaped how the workforce is distributed. As summarized by economist Elvira Nica, “As the cost of computing power has decreased, computers and robots have progressively taken the place of employees in carrying out unambiguous, codifiable undertakings.”8 More recently, the exponential rise of big data and data analysis have made more non-repetitive, dynamic, and large-scale tasks possible to computerize, as well.9 That is, decision-making is in part transferring from people to servers and neural nets. If this trend continues, there will be less and less work-based decision-making for people to take on which technology cannot itself perform. Additionally, if the rate of change in automation is exponential in nature as research indicates, eventually people may not be able to reskill at the rate that technology reshapes the workforce.10
Economically, automation makes sense: as Nica argues, “Insofar as digital technologies are employed for the intentions of surplus value production, they will generate results that are satisfactory for capitalist employers and detrimental for employees.”11 The scalable nature of tech means that a company can command a whole market with less labor and less redundancy. Amazon pushing operational efficiency onto its workers is a short-term market imperfection: these people should be paid comparably to do other, more enriching work and technology should eventually automate these functions. Whether in the short-term policy needs to counteract difficulties these workers face is a different (incredibly important) question.
3. Work plays a very important role in human psychological wellbeing
While at once automation promises to remove more drudging, dull aspects of work, so too is it increasingly overriding actions which make us feel autonomous. Law professor John Danaher warns even that “the automation of labor could rob us of meaningfulness…the necessary causal and mental link between our actions and the outcomes that are said to be constitutive of meaning.”12 Research and surveys have indicated, for example, that a large portion of the United States population would still work even if not economically compensated for their work. People want to feel like they are contributing members of society—to be a part of something bigger than themselves.13 To maintain this utility of work, whatever roles people take on must maintain this connection to society and this notion of contribution.
4. Life without work could be more meaningful than life with work
Research has indicated that much human satisfaction is derived from “flow-state-inducing activities” which can be obtained through work or through non-work.14 Much automation cuts out work with few longer-term implications (for example repetitive tasks) which are unlikely be flow-state-inducing. Moreover, much of the work we do is predictable and static which can be eliminated by smart automation.15 This is why automation largely to this point has been useful: it enables people to specialize in the activities they value the most. Down the line, if automation takes away certain flow-states in work, people should work to find similar flow in other areas such as their communities or families. People want to create and inspire regardless of the context. As a society, we are very focused on the notion of work but fail to assess how much of an opportunity cost it is relative to other meaningful pursuits.16
The first fundamental canon of the NSPE’s Code of Ethics is to “Hold paramount the safety, health, and welfare of the public.”17 Engineers implementing solutions which displace workers must ensure that these workers are indeed better off, either in new forms of work or outside of work. Investing in alternative forms of education and community programs are strong ways to open up opportunities for non-work flow-states and passions.
Having established (1) that incentives are aligned economically for the continued roll out of automation and (2) the potential moral upside to automation, we should continue to research it, implement it, and most importantly measure its progress. As mentioned, companies the likes of Amazon and Walmart have enormous societal impact in rolling out new tech. As they are the ones driving this innovation and, in the process, widening their competitive moats, they should be required to carefully manage transitions of labor.
A key consideration is equality: there will certainly be uneven deployment of automation between regions and companies.18 Amazon, for example, will benefit far more quickly from its own automation in the US than Jumia in Nigeria. Likewise, technological unemployment will funnel funds away from laborers towards managers, increasing income inequality.19 This threat to equality is why technology leaders have historically supported ideas the likes of universal basic income and automation taxes.20 In a society where not every individual is expected to work, we will need new models for distributing wealth and meaningful work which go beyond the fundamentals of capitalism.
Disclaimer: this was written for EAS 203: Engineering Ethics at the University of Pennsylvania. The assignment, titled “Communicating with the Public”, was to give my perspective on an emerging field of technology and its ethical implications.
1 Jacobs, Julian D. “Utilitarianism and the Market Economy.” Brown Political Review, Brown Political Review, 12 Apr. 2017, brownpoliticalreview.org/2017/04/utilitarianism-market-economy/..
2 Denise, Malcolm L. “Automation and unemployment: a management viewpoint.” The Annals of the American Academy of Political and Social Science 340, no. 1 (1962): 90-99.
3 Day One Team. “5 Facts to Know about Amazon Robotics.” EU Day One Blog, Amazon, 5 Sept. 2018, blog.aboutamazon.eu/innovation/5-facts-to-know-about-amazon-robotics.
4 Statista. “Net sales revenue of Amazon from 2004 to 2018 (in billion U.S. dollars)”, Statista, Statista, 2019, statista.com/statistics/266282/annual-net-revenue-of-amazoncom/.
5 Lieber, Chavie. “Emergency Calls Placed from Amazon Warehouses Depict Enormous Pressure Put on Workers.” Vox, Vox, 11 Mar. 2019, www.vox.com/the-goods/2019/3/11/18260472/amazon-warehouse-workers-911-calls-suicide.
6 Walmart. “Company facts.” Newsroom, Walmart, 2019, corporate.walmart.com/newsroom/company-facts
7 Chavie Lieber, “Emergency Calls Placed form Amazon Warehouses.”
8 Nica, Elvira. “Will technological unemployment and workplace automation generate greater capital-labor income imbalances?” Economics, Management and Financial Markets 11, no. 4 (2016): 68.
9 Elvira Nica, “Will technological unemployment.”
10 Danaher, John. “Will life be worth living in a world without work? Technological unemployment and the meaning of life.” Science and engineering ethics 23, no. 1 (2017): 41-64.
11 Elvira Nica, “Will technological unemployment.”
12 John Danaher, “Will life be worth living in a world without work?”
13 Morse, Nancy C., and Robert S. Weiss. “The function and meaning of work and the job.” American sociological review 20, no. 2 (1955): 191-198.
14 John Danaher, “Will life be worth living in a world without work?”
15 Parrish, Shane. “Smarter, Not Harder: How to Succeed at Work.” Farnam Street, 19 June 2018, fs.blog/2018/06/succeed-at-work/.
16 Nof, Shimon Y. “Automation: What it means to us around the world.” In Springer handbook of automation, pp. 13-52. Springer, Berlin, Heidelberg, 2009.
17 National Society of Professional Engineers, “Code of Ethics.” Ethics. NSPE, 2019.
18 Shimon Y. Nof, “Automation: What it means to us around the world.”
19 John Danaher, “Will life be worth living in a world without work?”
20 Morris, David Z. “Bill Gates Says Robots Should Be Taxed Like Workers.” Fortune. Accessed April 21, 2019. http://fortune.com/2017/02/18/bill-gates-robot-taxes-automation/.