(This article has been translated from French.)

In recent years, alarmist discussions of the threats of job automation have regularly been making the headlines. These fears can be explained by major advances in robotics and artificial intelligence (AI) that have enabled certain tasks to be automated. For example, robots can move around warehouses unguided and software interacts with customers to answer their questions. Other news items are reactions to the notions of a utopian, or dystopian, future presented by public figures like Bill Gates or Elon Musk. All this media attention stems from the possibility of a not-so-distant possible world in which the human workforce could be replaced by intelligent robots and algorithms that, while discreet, are frighteningly efficient.

In short, a “post-labour” world is presented as a possible – and sometimes even probable – scenario. A 2017 study added fuel to the fire of proponents of an inflationary interpretation of AI capabilities. Its authors predicted that 47 per cent of U.S. jobs face a high risk of being automated within the next two decades. But a closer look reveals that the most likely scenario in the near future is rather one of increasing inequities.

Since fall 2018, the Science and Technology Ethics Commission (CEST) – an advisory body of the Quebec government – has been studying AI’s effects on work and the ensuing social justice issues. The CEST published a whitepaper on this topic in fall 2019. It also published recommendations for public policy-makers on how to regulate AI deployment in the labour market in such a way as to promote social justice and the fair redistribution of the wealth generated.

Unlike the inflationary perspective described above, the CEST has adopted a “deflationary” view that is intended to be both realistic and to critically reflect on AI’s probable capabilities and consequences. This perspective is based in part on research showing that, while significant, the effects of AI on the labour market will hardly be the much fantasized post-labour world. On the contrary, the effect of AI technologies will likely amplify the outcomes of current dynamics. In other words, using this family of technologies in the workplace is likely to increase social inequalities and precariousness.

The impact of AI on work

Most of the available studies, including some specific to the Canadian labour market, do not expect AI-related technologies to lead to massive unemployment. Nevertheless, like other multi-pupose technologies, AI is apt to significantly affect how certain tasks are performed.

Hence, the unlikelihood of generalized unemployment should not make us overlook the more localized impacts that these technologies may have, such as job losses in specific fields, changes in certain tasks, shifts in in-demand skills, creation of jobs that benefit skilled workers, etc.

Are we moving toward greater unfairness and precariousness?

The deployment of AI could change the work prospects of the most vulnerable, especially people with less education, fewer skills and more difficulty entering and staying in the labour market. The jobs created by the use of AI could require advanced skills or competencies that these workers do not already possess or would find difficult to acquire. In addition, jobs filled by workers with lower-level skills may be more easily automated. New low-skilled jobs involving routine and time-consuming tasks may also add little value and pay a low wage.

In contrast, skilled workers are better positioned to change jobs or to benefit from the deployment of AI. Of course, they are not immune to the negative effects of AI, such as partial automation, which could make their jobs less rewarding or less diversified; the elimination of entry-level positions, which allow them to gain experience and move up; or the loss of income. However, compared with those with less education, skilled workers are more likely to experience increased productivity thanks to these technologies, and even to become wealthier. Investors and high-paid executives could also benefit from AI-based job automation.

We are therefore most probably heading toward more socioeconomic inequalities and more people experiencing poverty. The most immediate effect of AI is therefore not the advent of a post-labour world, but the deepening of the dynamics already at play in the knowledge economy.

What should be done about this risk of increasing inequality and precariousness? In collaboration with a committee of experts, the CEST has mapped out the best ways forward. The committee estimates that with the right modifications and upgrades, the current levers of wealth redistribution would be capable of mitigating the negative effects of AI in the workplace.

Wealth redistribution policies

In a society like Quebec, where social and fiscal policies ensure some redistribution of wealth, the current levers could go some distance in responding to the effects of AI. The CEST therefore does not recommend introducing a guaranteed basic income (GBI) universally distributed to all citizens. Knowing that a provincial expert committee has ruled out the idea of introducing an unconditional GBI, nothing leads us to believe that AI would radically change the equation and provide advocates with a decisive argument in favour of a GBI. This is why, rather, the CEST recommends that the government of Quebec introduce both universal and targeted measures to continue the fight against precariousness and socioeconomic inequality. In addition, the government must better address the situation of single people, who have been left behind by the anti-poverty programs of recent years.

Despite the robustness and adequacy of the current redistribution system, some adjustments will likely be necessary. For example, income support programs for the most disadvantaged should not be cut, as policy-makers may be tempted to do in order to balance the post-pandemic budget. Instead, these programs should be enhanced. For example, the CEST sees the work premium tax credit as a prime candidate for such improvement.

Job training

All workers – laid-off employment seekers or workers requiring retraining to retain their positions – must have access to the necessary job training. The CEST therefore recommends that the government expand the availability of and access to vocational training and continuing education. Currently, such access is unequal, as those most in need of continuing education tend to be the people who receive the least. In order for the deployment of AI technologies to be truly beneficial for all, the CEST recommends that the government of Quebec promote best practices to encourage workers to participate in the technological changes in their workplaces.

Better understanding of the impact of AI

It is imperative that the government gather and analyze data to better understand the current effects and to anticipate the future impacts of AI and automation on the workforce, particularly on lower-skilled workers. Equally important, the authorities responsible for the public subsidies and tax measures promoting AI development must analyze the impact of these investments.

The CEST is pursuing its analysis of the ethical issues of the effects of AI on the workforce and the organization of work, and is drafting a report on this subject. These reflections will address issues such as employment equity, employee welfare, worker autonomy, and the privacy and confidentiality of data collected over the course of employment. Profound changes are already underway in the workplace, but policy-makers have many tools at their disposal to help workers weather the storm. Let’s not get caught off guard!

Do you have something to say about the article you just read? Be part of the Policy Options discussion, and send in your own submission, or a letter to the editor. 
Jocelyn Maclure
Jocelyn Maclure chairs the Science and Technology Ethics Commission, serves as professor of philosophy at McGill University and is the Jarislowsky Chair in Technology and Human Nature.
David Rocheleau-Houle
David Rocheleau-Houle is a former ethics advisor at the Science and Technology Ethics Commission.

You are welcome to republish this Policy Options article online or in print periodicals, under a Creative Commons/No Derivatives licence.

Creative Commons License