Understanding AI's effect on jobs: Automation versus augmentation

Elsewhere, Stanford Professor Erik Brynjolfsson has identified the erstwhile Turing Test — a test of a machine’s ability to exhibit intelligent behaviour equivalent to that of a human — as “an instrument of harm” in creating an automation mindset for AI research at the expense of potential augmentation paths. Both Markoff and Brynjolfsson argue that it would be preferable if AI research travelled a more human-centric path, focusing on opportunities to augment rather than automate humans. Such AI applications would enable people to do things they could not previously do, creating a complementarity between such applications and human capabilities and skills.

They are joined in this belief by MIT Economist Daron Acemoglu, who has been vocal about the risks AI poses for job security unless more diverse research paths are chosen. Acemoglu sees the potential for AI in many sectors — from healthcare to entertainment. He also speculates on AI use in education. Current developments go in the direction of automating teachers — for example, by implementing automated grading or online resources to replace core human teaching tasks. But AI could also revolutionize education by empowering teachers to adapt their material to the needs and attitudes of diverse students in real time.

The advocates for human-centric AI list the mindset of AI researchers as the primary starting point for attitudes to change. Brynjolfsson argues, “A good start would be to replace the Turing Test, and the mindset it embodies, with a new set of practical benchmarks that steer progress toward AI-powered systems that exceed anything that could be done by humans alone.” The underlying hypothesis is that if the objectives of AI research are changed, this would steer the economy away from the potential loss of jobs, devaluation of skills, inequality and ensuing social discord. In this way, society could avoid what Brynjolfsson calls the “Turing Trap,” whereby AI-enabled automation leads to a concentration of wealth and power.

In a recent paper, we question this hypothesis, asking whether it is really the case that the current technical objective of using human performance of tasks as a benchmark for AI performance will result in the negative outcomes described by Acemoglu and colleagues. Instead, we argue that task automation — especially when driven by AI advances — can enhance job prospects and potentially widen the scope for employment of many workers.

The neglected puzzle piece we highlight is the potential for changes to the skill premium: AI automation of tasks exogenously improves the value of the skills of many workers, expands the pool of available workers to perform other tasks and in the process, increases labour income and potentially reduces inequality. We label this possibility the ‘Turing Transformation.’ The distributional effects of technology depend more on which workers have tasks that get automated than on the fact of automation per se. This defines a Turing Transformation. What is happening is that AI involves a task that requires specialized skills, and the automation of that task opens up opportunities for more workers. In effect, workers with generic skills are helped when AI is adopted because they are able to participate in jobs previously only available to those with specialized skills.

However, suppose the only workers are skilled workers. Under these assumptions, used by Acemoglu, if there are large economies of scope or AI involves a high unit cost, wages would fall if AI were adopted. This is the situation that one might characterize as a Turing Trap. In this model, an AI that is built with the intention of replacing a human in a task — that is, an automation mindset — turns out to be augmenting for the majority of workers because it opens up an opportunity to work on other tasks that would previously have been bundled as a job created for relatively scarce workers. In this model, more workers compete with one another, but the productivity effect is such that total labour income rises.

Broadly speaking, the implication is that the notion that automation and augmentation involve distinct mindsets with distinct outcomes for workers misses some relevant features. Different workers have different skills, and many of the developments in AI with the potential for widespread impact entail replicating an aspect of the intelligence of a small number of higher-wage workers. In doing so, the technology could create opportunities for a much larger number of workers, enabling new opportunities for employment, along with the potential for higher wages and career choice.

When considering automation versus augmentation, the heterogeneity of worker skills is fundamental. One worker’s automation is another’s augmentation. Automation of rare high-value skills can mean augmentation for everyone else. Similarly, augmentation that complements the lucky humans with rare high-value skills can mean increased inequality and a hollowing out of the middle class.

Another potential criticism of this perspective is that it is not always obvious whether a technology replaces something that is currently a human skill, and thus the line between augmentation and automation can be blurry. We take the distinction as given. If it is blurry whether a technology is ‘intelligence augmentation’ or ‘human-like artificial intelligence,’ that enhances our broader point that this distinction is not useful.

Also read: AI in marketing: How customer-centric companies can benefit from data privacy regulations

Examples of the Turing Transformation

The discussion of automation and augmentation has a new urgency because of advances in AI over the past decade. These advances are primarily in a field of AI called machine learning, which is best understood as prediction in the statistical sense. By prediction, we mean the process of filling in missing information.

Our examples will focus on advances in prediction technology, though our broader point about the value of automation versus augmentation is not specific to prediction machines. Technologies that replace the core skills of some workers can enable others to get more out of their skills.

There is already some evidence that AI might be particularly likely to affect the tasks performed by high-wage workers. Webb found that the most common verbs in machine learning patents include "recognize," "predict," "detect," "identify," "determine," "control," "generate" and "classify." He also finds that these verbs are common in tasks done by relatively high-wage workers. It is an open question whether automating these tasks will simply reduce the wages of those who are already doing well or whether it will create new opportunities for lower-wage workers. But our model suggests that automation has the potential to reduce inequality, not just by making those with higher wages worse off but by creating a Turing Transformation for many more workers. Following are some examples.

PERSONAL TRANSPORTATION: Since 1865, London taxi drivers have had to pass a test demonstrating mastery of ‘The Knowledge’ of the map of the city’s complicated road networks. Most drivers had to study three to four years before passing the test. This is a skilled occupation, requiring incredible memory skills and the discipline to spend the time studying. Fifteen years ago, no one could compete with the ability of London taxi drivers to navigate the city. Today, the taxi drivers’ superpower is available for free to anyone with a phone. Digital maps mean that anyone can find the best route, by driving, walking or transit, in just about any place in the world. The mapping technology substitutes for the driver’s navigation skill. It doesn’t provide something new, but it replicates a human skill more cheaply. As a result, taxi driver wages have fallen. This is precisely what Markoff, Brynjolfsson et al. warn against.

Automation of the taxi drivers’ competitive advantage, however, has created opportunity for millions of others. By combining navigation tools with digital taxi dispatch, companies like Uber and Lyft have enabled almost anyone with a car to provide the same services as taxi drivers.

Put simply, technology automated the core skill that only a handful of skilled humans could already do. And in the process, it provided the opportunity for many without those skills to work in the same industry. In the U.S., there were approximately 200,000 professional taxi and limo drivers in 2018. Today, more than 10 times that number drive for Uber alone.

CALL CENTRES: Around the world there are millions of customer service representatives working in call centres where productivity is carefully measured in terms of calls per minute and satisfied customers. Like other industries, worker productivity is uneven. The most skilled agents are much more productive than the median, and new workers improve rapidly over the first few months.

A recent paper by Brynjolfsson and colleagues looks at the deployment of AI in a software support call centre. These calls are relatively complicated, averaging over 30 minutes and involving the troubleshooting of technical problems. The AI provides real-time suggestions on what the call centre worker should say, and the worker can choose to follow the AI or ignore it. The most productive workers benefit very little, if at all from this. They may even rationally ignore the AI’s recommendation. In contrast, it is the least productive workers and the newer workers that benefit. Notably, their relative productivity compared to the most productive workers increases: The AI reduces the gap between the less skilled and more skilled workers.

The paper provides suggestive evidence that this is because the less-skilled workers learn what their more skilled peers would do in a given situation. This technology is automation as defined by Markoff: It involves machines that do what humans do, rather than machines that do something that humans can’t do. It is used as decision support, and therefore can be seen to serve as a complement to all of the human workers, regardless of their skill. In practice, however, the technology helps the least skilled.

MEDICINE: A large and growing body of research is showing the potential for AI to provide medical diagnoses. Underlying this research is the insight that at its core, diagnosis is prediction: It takes information about symptoms and fills in missing information of the cause of those symptoms. Diagnosis, however, is a key human skill in medicine. Much of the training doctors receive in medical school — and the selection process they go through in order to get into medical school — focuses on the ability to diagnose.

Other workers in the medical system may be better at helping patients navigate the stress of their medical issues or providing the day-to-day care necessary for effective treatment. Diagnosis is perhaps the central skill that sets doctors apart. An AI that does diagnosis automates the task requiring that relatively rare skill. It is not augmented intelligence but a replacement for human intelligence.

In 2021 there were 760,000 jobs for physicians and surgeons in the U.S., earning a median income of over $200,000 per year. Automating the core skill that many of them bring to the table could eliminate much of doctors’ value, even leading to stagnating employment and wages. Again, exactly the worry that Brynjolfsson and Markoff warn against when AI replicates human intelligence.

However, in 2021 there were also three million jobs for registered nurses and millions for other medical professionals, including pharmacists, nurse and physician assistants and paramedics. While AI diagnosis would likely negatively affect many doctors, if these other medical professionals could perform AI-assisted diagnosis then their career opportunities, and possibly wages, could increase substantially.

LANGUAGE TRANSLATION: Another task currently performed by skilled workers that AI could take over is language translation. Many people speak multiple languages, and in many workplaces this ability confers an advantage. Speaking French and English is an advantage in many Canadian workplaces, particularly for the hundreds of thousands who work in the civil service or in regulated industries. Similarly, people who speak multiple languages have an advantage in many international business opportunities. Of course, many people work as translators, earning their income directly from their ability to translate between languages. For written texts, when the goal is simply to communicate with little regard for eloquence, AI is already good enough to replace many human translators. For large-scale translations and real-time translation of verbal communication, there are reasons to expect machine translation to be good enough to deploy commercially in the very near future.

These advances are probably bad news for the tens of thousands of language translators in the U.S. However, they are likely good news for many others. Brynjolfsson et al. report that AIs used for translation enhance the capacity of sellers on eBay, increasing exports by 17.5 per cent. AI that automates language translation enables enhanced communication across the world. It likely means more trade, more travel, faster integration into workplaces for recent immigrants, more cross-cultural exchange of ideas and perhaps even different social networks. Those whose jobs have been constrained by an inability to speak or write in multiple languages would no longer face those constraints.

WRITING: The ability of AI to write goes beyond translation. When OpenAI released ChatGPT in November 2022, it quickly gained millions of users due to its ability to produce well-written prose on a wide variety of topics. It can even produce high-quality five-paragraph essays, leading to worries about the future of take-home exams and the potential for widespread cheating. It can write eloquent emails, longer articles and summarize research and news events. Because summarizing, interpreting and writing is such an important part of knowledge work, MIT Economist Paul Krugman worried that ChatGPT means that “robots are coming for the skilled jobs.”

Summarizing and writing are clearly not cases of a machine doing something that is beyond human capability. As such, this is automation, not augmentation. Or in Markoff ’s language, it is ‘artificial intelligence for duplicating human behaviour’ — not intelligence augmentation that attempts to expand human abilities.

That, however, depends on the human. Many people don’t write well. With ChatGPT, they can quickly compose communications to customers, suppliers or friends without fear of grammatical mistakes and without the need to stress about how to get their ideas down on paper. This could enable millions of people to benefit from skills other than writing. When almost anyone has the ability to write clearly, there will be changes to who is capable of which jobs, with many people in the bottom half of the current income distribution receiving new opportunities while some at the top face enhanced competition.

As with taxi drivers, those who make their living writing will be affected positively and negatively. They may become more efficient, given that AI can summarize articles and write or revise drafts. But they will also face more competition for their work and, like taxi drivers, their wages may fall as their skills are no longer scarce.

Also read: Human intelligence must guide artificial intelligence

Intelligence Augmentation and Inequality

We will now provide examples of information technologies that are best seen as intelligence augmentation under Markoff ’s definition — as technologies that do things that are not possible for humans to do. In this sense, they are outside the motivating model, as they do not involve directly automating a specific task done by a human worker — although, as we have emphasized, one person’s augmentation could be another’s automation.

COMPUTERIZATION: As Brynjolfsson has put it, “computers are symbol processors.” They can store, retrieve, organize, transmit and transform information in ways that are different from how humans process information. Markoff notes that modern personal computers have their root in Douglas Engelbart’s augmentation tradition. Unlike AI, which we believe may decrease inequality, computerization increased inequality and led to polarization of the U.S. wage distribution, expanding high- and low-wage work at the expense of middle-wage jobs.

This is because, while some tasks done by computers could be done by humans, much of the changes are a result of complementarity between the skills of the most educated workers and the identification of new ways to use the machines. In other words, rather than directly replacing a task done by middle income workers — as AI does — computers complemented the skills of those already near the top of the income distribution, thereby increasing their productivity for tasks that were already being done by humans.

Again, quoting Brynjolfsson: “As computers become cheaper and more powerful, the business value of computers is limited less by computational capability and more by the ability of managers to invent new processes, procedures and organizational structures that leverage this capability.” Census data comparing business software investment with employee wages shows that within and across firms, software investment increases the earnings of high-wage workers more than that of low-wage workers. Computers displaced the workers performing routine technical tasks in bookkeeping, clerical work and manufacturing, while complementing educated workers who excel in problem-solving, creativity and persuasion.

DIGITAL COMMUNICATION: The Internet represents another technology that does something distinct from what humans can do. For the most part, as Markoff notes, it does not replace specific tasks in human workflows. Instead, it allows computers to communicate with each other, sending information between millions of devices. This information is a complement to the human skills of interpreting and acting on information.

People and places at the top of the income distribution benefited from the technology while those with less education benefited less. To the extent that there are differences between augmentation and automation technologies, the Internet is more of an augmentation technology. As such, it complemented the skills of those who were already at the top of the distribution.

The above discussion warrants an important caveat: Many have called computerization and digital communication "automation." Formally, it is difficult to classify technologies as automating or augmenting, and we do not want to take a strong stand on which technologies belong in which category. That is an aspect of our underlying point: One person’s augmentation is another’s automation. What matters is the distribution of workers whose skills are complemented.

The first 50 years of computing introduced many technologies that appear to be intelligence augmenting, creating new capabilities and new products and services. The last 10 years have seen a rise in AI applications, whose inventors directly aspire to automate tasks currently performed by humans. On the surface, technologies labelled as "augmentation" appear to complement human workers, while automation technologies appear to substitute for human workers. Therefore, many scholars have called for engineers, scientists and policymakers to focus on augmentation technologies over automation.

An important aspect of this argument is the idea that complements to human labour will reduce income inequality, while substitutes for human labour will increase it. We argue that this dichotomy is misleading. A key aspect of understanding the impact of intelligence technology on inequality and the well-being of most workers is the heterogeneity of workers’ skills. Our bottom line is this: A technology that directly substitutes for rare and highly-valued skills could create enormous opportunities for most workers.

In Closing

With technological change, the winners and losers are not determined by whether the technology seems to replace or augment human tasks. Instead, winners and losers are determined by whether the augmentation affects lower-wage workers and whether automation affects those who are already doing well.

Perhaps the best targets for computer scientists and engineers looking to build new systems is not to find intelligences that humans lack. Instead, it is to identify the skills that generate outsized income and build machines that allow many more people to benefit from those skills. As noted herein, this may be what is already happening with AI that recognizes, predicts, determines, controls, writes and codes.