From Adoption to Embrace of Generative AI in the Workplace: A Dialogue Between Educational Sciences and Psychology

October 14, 2026

Integrating AI into organizations is about more than just adopting the technology. Through the combined perspectives of Joseph Baud-Grasset and Stéphane Bonzon, the third installment of this series of articles on the topic explores the keys to thoughtfully integrating AI into everyday operations.

Chapter 3: AI in the Workplace: Learning, Adapting, and Preserving the Meaning of Work

The integration of artificial intelligence into organizations does not depend solely on professionals’ ability or willingness to adopt new tools. It raises broader questions about the conditions under which these transformations are implemented: time available for learning, support, peer-to-peer exchanges, cognitive load, and the ability to maintain autonomy in their work practices. Through the contrasting perspectives of Joseph and Stéphane, this article explores the conditions necessary for an adoption of AI that is sustainable, thoughtful, and compatible with the quality of work and the well-being of professionals.

Starring:

Joseph Baud-Grasset – Training Coordinator & Quality Manager

Stéphane Bonzon – Career Counselor

Do the challenges encountered with AI stem primarily from individuals or from the conditions under which work is organized ?

Joseph :

I tend to be wary of explanations that place the blame primarily on individuals. When it’s difficult to master something, the first question isn’t always whether people are motivated or competent, but rather under what conditions they need to learn.

We often see a kind of paradox: organizations ask professionals to quickly adopt new tools while maintaining the same workload, the same deadlines, and sometimes the same performance metrics. As a result, employees are expected to learn without being given adequate time to do so.

However, learning takes time. Testing a tool, understanding what it actually offers, adjusting one’s practices, or sharing experiences with colleagues are all part of the learning process. If these activities remain unseen or are perceived as a waste of time, it automatically becomes more difficult to adopt the tool.

Of course, there are individual differences in attitudes toward innovation or change. But what we have long known in the field of education is that behaviors cannot be separated from their context.

The issue, therefore, is not only about the preparation of professionals, but also about the ability of organizations to create conditions conducive to learning.

When these conditions are not present, what is sometimes interpreted as resistance is often a perfectly rational form of protection against an additional burden.

Stéphane:

Joseph rightly points out that learning depends on the organizational context. The Job Demands–Resources model (Bakker & Demerouti, 2007) allows us to ask a simple question about the use of AI in the workplace: Is AI a resource or an additional demand? Depending on how it is introduced, it can be both at the same time. It assists with certain tasks while increasing the cognitive load due to new demands, such as verifying its outputs or managing its integration into a workflow. If we measure only the gains, we fail to see the silent increase in such cognitive demands.

What specific conditions would enable a healthier and more sustainable adoption of AI in organizations ?

Joseph:

If I had to identify a few priorities, I would start with a simple idea: to view the adoption of AI as a learning process rather than as a technology deployment project. At the very least, the two should not be separated.

First and foremost, this means setting aside time—not just for training, but for experimenting in real-world work situations, comparing different approaches, adjusting practices, and learning from experience.

This, in turn, involves supporting collective dynamics. Peer-to-peer exchanges, communities of practice, and internal mentors often play a more important role than top-down, highly academic training programs. It is these spaces that enable individual efforts to be transformed into shared learning.

Another essential condition is recognizing that adoption is necessarily heterogeneous. Not everyone progresses at the same pace, nor do they all have the same needs or uses. Seeking to standardize practices too quickly risks producing superficial compliance rather than genuine learning.

Finally, I believe it is important to make room for reflection. In a world where everything is moving faster and faster, professionals need opportunities to examine the impact of these tools on their work, their skills, and the quality of their work.

The challenge is not simply to use AI more. The challenge is to build a relationship with AI that preserves the foundations of professional work, such as judgment, a sense of responsibility, cooperation, and the ability to learn.

Stéphane:

To the pedagogical conditions set forth by Joseph, I would like to add two conditions that are not technical in nature. The first is to legitimize non-use. In an environment where the use of AI tends to become second nature, non-use can be one of the strongest expressions of discernment. This is not because rejecting AI would be preferable in and of itself, but because certain situations call for restraint. Thus, an informed use of AI requires that a professional be able to say they will not use it in a given situation, without this being interpreted as resistance, but rather as a legitimate and relevant expression of expertise. Added to this is the essential consideration of the mental load—more or less visible and growing—induced by artificial intelligence and the new cognitive demands of its development (see, for example, my recent article for the ISFB Observatory, Bonzon, 2026). The overall goal here is to move toward an AI that truly serves humans—that enhances them, one might say, but under no circumstances diminishes them.

If each of you were to identify one key area of concern and one priority for action for organizations, what would they be ?

Joseph:

If I had to highlight one point to watch out for, it would be the temptation to confuse the speed of adoption with the quality of adoption. Today, many organizations measure the number of activated licenses, usage rates, or productivity gains. These metrics are useful, but they don’t tell us much about what professionals are actually learning or how their practices are evolving.

My priority, therefore, would be to create learning environments rather than simply implementing programs. This means setting aside time for experimentation, supporting peer-to-peer exchanges, recognizing trial and error as a normal part of the learning process, and accepting that not everyone progresses at the same pace.

Ultimately, the question is not just how to integrate AI into organizations. The challenge is how to enable professionals to use these tools to develop more thoughtful, relevant, and autonomous practices. It is this capacity for learning that will, in the long term, constitute organizations’ true strategic resource—one that can be measured in professional practices.

Stéphane :

In the workplace, AI raises issues related to occupational health in the strict sense of the term. This aspect of health—which concerns us all—seems to me to be measured, in particular, by the strength of our sense of competence, by our ability to continue advocating for what we consider to be quality work, and by our capacity to

Bibliography:

• Bakker, A. B., & Demerouti, E. (2007). The job demands–resources model: State of the art. Journal of Managerial Psychology, 22(3), 309–328. https://doi.org/10.1108/02683940710733115
• Bonzon, S. (April 10, 2026). Generative AI and Work: Beyond the Hype, What Research Already Reveals. ISFB Observatory on Banking and Financial Skills.

© Institut Supérieur de Formation Bancaire (ISFB). All rights reserved.
The analyses and content published by the ISFB may be quoted or reproduced in part, provided that the source is clearly mentioned. Any full or substantial reproduction of this article in another medium or format is subject to the prior written authorization of the ISFB. In order to facilitate reading and without any intention of discrimination, the masculine gender is generally used, in accordance with the grammatical rule that allows it to be used as a neutral value to refer to a group of people comprising both men and women. This publication is intended for ISFB members and their employees in Switzerland, as well as anyone interested in finance in Switzerland. It is not intended to be read or distributed in any jurisdiction where its distribution would be prohibited.

How can we promote the sustainable adoption of artificial intelligence in the workplace? Discover the learning, support, and autonomy needed to maintain the quality of work and the well-being of employees.