AI in Quebec: 5 tipping points to avoid losing control

1. Introduction: The awakening of collective intelligence

Quebec is experiencing a fascinating period of strategic dissonance: while we are among the world leaders in artificial intelligence, we are suddenly realizing that this technological advantage could slip through our fingers if it is not accompanied by firm ethical direction. The urgency is no longer merely scientific; it is civilizational. This shift crystallized in March 2023, when Professor Yoshua Bengio and other leading international figures sounded the alarm, calling for a moratorium on advanced systems.

In response, the government entrusted the Quebec Innovation Council with an unprecedented mandate: to conduct a non-partisan , transparent, and inclusive collective reflection. This monumental undertaking mobilized 250 experts, gathered 420 citizen contributions, and brought together 1,500 participants at a public forum. This is no longer the work of a few isolated researchers, but rather the operationalization of a national consensus to define how our society should coexist with machines that now mimic our most complex cognitive abilities.

2. The end of the digital “Wild West”: Towards a Quebec framework law

The era of informal self-regulation is coming to an end. The "Ready for AI" report is unequivocal: the laissez-faire approach of recent years has generated unacceptable risks of discrimination, surveillance, and disinformation. The priority recommendation (RP-1) therefore advocates for the creation of a Quebec framework law dedicated to AI .

Unlike rigid approaches, this framework proposes a gradation based on the severity of the risks posed by the systems. Rooted in the principles of the Montreal Declaration , this legislation would aim to:

• Protecting fundamental rights against high-risk systems, such as health or financial regulations.

• Guarantee transparency , including the explicit right to know whether one is interacting with a human or an algorithm.

• Establishing accountability between designers and users to prevent harm.

• Establish an agile governance structure , capable of iterating in line with technological breakthroughs.

As Luc Sirois, Quebec's Chief Innovator, states: "This framework must reflect our values to better protect human rights and freedoms while simultaneously fostering innovation."

3. Work will never be the same again (and our laws must change)

AI is not just a productivity lever; it's a social paradox that must be carefully navigated. On the one hand, it offers a solution to address labor shortages and free humans from repetitive tasks. On the other hand, it carries the risk of devaluing skills and leading to dehumanizing algorithmic management.

Recommendation RP-3 calls for a profound modernization of labour law and social policies. To ensure the social acceptability of this transition, Quebec must avoid "technological unemployment" by giving workers a real voice. Dialogue between employers, unions, and training institutions is essential to transforming the threat of automation into a sustainable retraining strategy. The challenge is to protect workers' dignity while seizing the opportunity for enhanced human-machine collaboration.

4. Protecting the Soul of Quebec: Culture and Democracy Tested by Algorithms

Quebec democracy faces systemic threats: deepfakes and automated disinformation can now saturate the public sphere at a speed exceeding our capacity to react. At the same time, our culture is suffering a visibility crisis. On platforms like Netflix, 70% of viewed content comes from opaque algorithmic recommendations that tend to ignore Francophone and Indigenous productions.

To counter this erosion, Recommendation RP-9 proposes the creation of a national cultural database . Drawing inspiration from the Swedish model , Quebec aims to train its own models on high-quality local data. The objective is both political and related to identity: to ensure that AI speaks our language and reflects our narratives, rather than relying exclusively on models trained on foreign data that ignore our specific characteristics.

5. Frugal AI: The environmental paradox

It's time to confront the materiality of the immaterial. While Quebec has an advantage with its hydroelectric power, the material reality of AI remains significant: massive water consumption for servers, growing energy needs, and soil contamination linked to rare metal extraction. Quebec must become a champion of frugal AI . This implies assessing the environmental footprint across the entire lifecycle of the system, so that the technology's climate promises are not negated by its own voracious appetite for critical resources.

6. Becoming a model: The Quebec state as a laboratory

The government cannot limit itself to a policing role; it must be an exemplary and transparent user. The "AI for Quebec" program (RP-11) aims to operationalize AI in essential government missions (health, education, climate) with increased rigor.

This ambition involves creating a public registry of AI systems used by the administration, including risk analysis tools inspired by the federal government's algorithmic impact assessments. This commitment to transparency aligns with Yoshua Bengio's vision: "In reality, we don't have to choose. We can have progress in AI and security and democratic participation."

7. Conclusion: A project that has only just begun

Quebec is at a crossroads. The roadmap has been laid out, but success will depend on its execution. The creation of a transitional steering committee (RP-12) within the next three months will mark the shift from reflection to legislative and budgetary action.

We have the opportunity to shape a technology that reflects who we are.

However, one question remains for every citizen, creator and decision-maker: what part of our autonomy are we prepared to delegate to algorithms, and at what cost to our collective future?