AI and employment in Quebec: Beyond statistics, a profound transformation of work 10 minutes

The Institut du Québec (IDQ) has just published an in-depth study on the impact of artificial intelligence on employment in Quebec. This publication was the subject of an article in La Presse on Wednesday, January 15, with the headline “L’intelligence artificielle menace 810 000 emplois au Québec” – a somewhat alarmist angle that deserves to be nuanced and contextualized.

This study comes at just the right time to enrich our understanding of the transformation of work by AI. Last fall, I analyzed Statistics Canada data revealing that around 60% of Canadian jobs would be exposed to AI, while highlighting the strong potential for complementarity for almost half of them. The IDQ study now offers a different perspective and introduces a new concept: that of “vulnerable jobs”.

Where Statistics Canada assessed AI exposure (the degree to which a job could be transformed by AI) and complementarity (the potential for synergy between humans and AI), the QDI looks at vulnerability from a broader angle: that of workers’ ability to adapt and reorient themselves professionally in the face of these transformations. This change of perspective is worth considering.

As I read through the study, I immediately wondered about that term “vulnerability” that keeps cropping up in their analyses. What does it really mean to be “vulnerable” to automation, and how does this approach differ from simply measuring exposure to AI?

To give readers an overview of the potential impacts of this transformation, I’ve grouped together all the key statistics presented in the IDQ study at the end of this article. These figures, which cover sectors of activity as well as demographic groups and levels of education, help us to better grasp the scope and complexity of this transformation that is taking place in Quebec.

The IDQ concept of “vulnerability”

What exactly makes a worker or a job “vulnerable” according to the IDQ?

The answer is more complex and interesting than it seems. The IDQ proposes a two-dimensional definition of vulnerability, which goes far beyond the simple risk of automation. According to them, to be considered vulnerable, a worker must face a double challenge:

• For the first dimension of this vulnerability, the IDQ has drawn on the widely recognized work of Frey and Osborne (2013), two Oxford researchers who have developed a methodology for assessing the probability of job automation. According to this approach, a job is considered “at risk” when it has a probability of more than 70% of being automated in the foreseeable future. In particular, these researchers have identified three types of skills that are naturally resistant to automation: social intelligence, creativity, and the ability to perceive and manipulate the physical environment. But in the IDQ study, this risk of automation is only the first part of the equation.
• The second aspect, and this is where the analysis becomes particularly novel, concerns the worker’s ability to transition to another job that is less threatened. Drawing on the work of the OECD, the IDQ considers a transition to be “acceptable” only if it meets strict criteria: no more than six months’ training required, a pay cut limited to a maximum of 10%, similar skills required, and a common area of expertise with the current job. If a worker whose job is threatened cannot easily make such a transition, he or she is considered vulnerable.

This rather precise definition helps us understand why the QDI arrives at different population proportions than Statistics Canada. It’s not simply a question of who will be affected by AI, but rather of who risks finding themselves at a professional impasse in the face of this transformation.

It’s important to stress a crucial point here: talking about “vulnerable” jobs or workers doesn’t mean that these jobs are doomed to disappear, or that these workers will inevitably find themselves out of work.

Instead, the IDQ study presents us with a risk map, a kind of early warning system that enables us to identify where to focus our adaptation and training efforts..

The emerging reality will certainly be more nuanced than the statistics suggest. The history of previous technological revolutions has shown us time and again that predictions of massive job losses have rarely come true. Jobs evolve and transform, and new opportunities emerge. A job considered “vulnerable” today could very well reinvent itself tomorrow, integrating AI as a tool rather than a replacement.

Two distinct paths for AI deployment

The IDQ study presents a particularly enlightening perspective on how AI can be deployed in organizations. Rather than seeing automation as a uniform phenomenon, it distinguishes two fundamentally different paths, illustrated in the chart below (in french):

Complementary AI: Amplifying rather than replacing

The first path, that of complementary AI, aligns perfectly with what I described in my previous article as the roles of digital “Colleague” and “Coach”. In this approach, AI becomes a tool that amplifies human capabilities rather than replacing them. This path manifests itself in two main ways according to IDQ:

1. Supporting the existing workforce: AI acts as an assistant that enables workers to be more efficient and focus on higher value-added tasks.
2. The creation of new jobs: the integration of AI generates new needs and therefore new professional opportunities, whether in the development, implementation or advanced use of these technologies..

AI substitution: Task transformation

The second path, that of AI substitution, corresponds more to what I called the role of digital “Clerk”. It aims to automate certain tasks previously performed by humans. However, the IDQ study reminds us that this substitution :

• Is rarely total for a given job
• Generally takes place gradually
• Can be particularly relevant in a context of labor shortage

The importance of choosing the right path

What makes this distinction particularly important is that it underlines the fact that the impact of AI on employment is not predetermined. It largely depends on the choices organizations make in how they deploy these technologies. A company that favors complementary AI will probably need to invest more in training and coaching its employees, but will potentially be able to create more value by combining the strengths of AI with human expertise.

This duality of approaches also helps us to better understand why some sectors or professions seem more “vulnerable” than others. It’s not so much a question of technological sophistication as of the ease with which tasks can be either complemented or substituted by AI.

In my experience with organizations adopting AI, those that are most successful are often those that start with the complementary approach. They use AI to strengthen their existing teams before considering any form of substitution, creating an environment more conducive to innovation and adaptation.

Training as a bridge to the future to facilitate professional mobility

This more detailed understanding of vulnerability enables us to better target the necessary interventions. Where my previous article talked about the importance of “AI literacy” and critical thinking, the IDQ study suggests that organizations need to go further. It’s not just a question of understanding AI or knowing how to use it, but of developing skills that facilitate professional mobility.

IDQ’s recommendations reflect this deeper understanding. It’s not just a question of training people in AI, but of creating accessible transition “bridges”: short but effective training pathways, cross-disciplinary skills development programs, and prior learning recognition mechanisms that facilitate professional transitions.

While my previous article stressed the importance of “AI literacy” and critical thinking, the IDQ study goes further, proposing concrete recommendations for the education system and continuing education. In particular, it highlights the urgent need to:

• Rethinking training support measures for people in employment
• Developing a more agile continuing education offering
• Systematically integrate the impact of technology into training planning

A study that sheds light on the path to transformation

A closer look at this IDQ study reveals a considerable departure from the alarmist message that has been making the headlines. Beyond the figures on “vulnerable” jobs, the study actually offers us a valuable tool for understanding and navigating the transformation that is taking place in our organizations.

The introduction of the concept of “vulnerability”, which goes beyond mere exposure to AI to consider workers’ ability to adapt, helps us to better target our interventions. This approach enables us to see where to focus our efforts, whether in terms of training, skills development or career reorientation.

The distinction between complementary and substitute AI also reminds us that we have a choice in how we integrate these technologies. Organizations can opt for an approach that strengthens and enhances the capabilities of their employees, rather than simply seeking to automate jobs.

For business leaders, this study offers a valuable framework for reflection. It identifies the sectors and jobs that will require particular attention, while suggesting avenues for a thoughtful and responsible adoption of AI. It’s not so much a question of rushing into automation as of planning a transformation that benefits both the organization and its employees.

Ultimately, the IDQ study invites us to see AI-driven transformation not as an imminent threat, but as an opportunity to rethink our approaches to work and professional development. It is by understanding these nuances and possibilities that we can truly take advantage of this technological revolution, while ensuring that it benefits as many people as possible.

Key statistics from the IDQ study: A numerical portrait of transformation

Volume and range

• 810,000 Quebecers work or are looking for work in occupations vulnerable to automation
• This represents 18% of the Quebec workforce.
• 96 professions have been identified as vulnerable to automation

Impact by business sector

• Manufacturing: 142,465 vulnerable workers (the largest number)
• Retail: 118,425 vulnerable workers
• Accommodation and food services: 116,475 vulnerable workers
• Health care and social assistance: 72,220 vulnerable workers
• Construction: 70,665 vulnerable workers

Demographic impact

• Young people aged 15-24 account for 13% of the workforce but 24% of vulnerable workers
• 27% of adults over 25 without a diploma are in vulnerable employment
• Only 8% of university graduates are in vulnerable employment
• Women represent 48% of the workforce but 52% of vulnerable workers

Impact by level of education

• High school graduates (DES): 27% of vulnerable workers
• Holders of a Diploma of Vocational Studies (DVS): 22% of vulnerable workers
• College diploma (DEC) holders: 14% of vulnerable workers
• Holders of a bachelor’s degree: 8% of vulnerable workers.ailleurs vulnérables

Impact by professional category

• 59% of jobs in manufacturing and utilities are vulnerable
• 47% of jobs in natural resources and agriculture are vulnerable
• 25% of jobs in sales and services are vulnerable
• 27% of jobs in business, finance and administration are vulnerable

Technoloy adoption

• Only 12% of Quebec companies plan to use AI to produce goods or deliver services in the next year
• Large companies more likely to adopt AI than SMEs
• Professional services and finance sectors most advanced in AI adoption

---

Moov AI uses generative AI to create this blog. 
We used generative AI to speed up the production of this blog. For the text, a multi-stage process was followed. NotebookLM was first used to analyze and extract key points from the two studies (Statistics Canada and L’institut du Québec). Claude AI was then asked to use its ‘projects’ functionality to enrich the document base, including the studies, NotebookLM’s comparative analysis and reference articles. An initial exchange with Claude AI enabled us to explore different possible structures for the article. Once a plan had been chosen, an in-depth conversation with Claude AI guided the writing of the different sections, making sure to maintain consistency with the style of the previous articles provided as reference. Mathieu then added his personal touch to the text.

DeepL’s “translate text” function was used to translate this text from French into English.

For the image, the marketing team used MidJourney to generate the header image using this message : « An abstract illustration used for a header of a blog about a framework for assessing the quality of artificial intelligence systems. Use overlapping translucent geometrical shapes symbolize the interconnected, multi-faceted nature of the framework. Use simple graphics. The mood is bright and optimistic. –ar 2:1 »

---