Need for Safe and Productive Development and Use of Artificial Intelligence

Honourable senators, on this World Book and Copyright Day, I rise to speak to Senator Moodie’s inquiry on the safe and productive development and use of artificial intelligence in Canada.

First, I want to thank Senator Moodie for initiating this inquiry, which today touches on all aspects of human activity to a degree and at speeds which we are now beginning to grasp. In fact, given the exponential pace of AI’s evolution, I fear that some information in my speech could be out of date by the time I finish. This should give you some idea of the urgency of addressing this question.

During this inquiry, a number of our colleagues have discussed issues such as security, innovation, regulation, rights and sovereignty. Others have taken positions in favour of AI culture, and I thank them for that. With this in mind, I want to share my thoughts, which I am sure will gain from further study in the months to come.

Vaclav Havel, Czech playwright, essayist and statesman, wrote that culture is the conscience of the nation. If that is true, and I believe it is, then the technological revolution transforming culture deserves the full attention of the Senate of Canada and proactive intervention by the Canadian state.

Recognizing that AI is already transforming our relationship with culture, language and creativity — in other words, our relationship with human expression — the fundamental question is this: What becomes of a society and its cultural sovereignty when technological advancements tend to shape its memory, its narratives, its aesthetics, its linguistic diversity and its collective imagination?

AI does not merely affect our jobs and markets. It influences what we see, hear, read and create, and ultimately, what we pass on to others. The debate around AI goes far beyond technical and commercial considerations. It is cultural, democratic and profoundly human.

Yes, it is also economic.

Colleagues, let us remember that Canadian culture contributes more than $65 billion to our GDP and supports nearly 700,000 jobs. It generates more employment output per dollar than several sectors often regarded as foundational, including oil and gas, manufacturing and agriculture. We should never lose sight of that.

Artificial intelligence generates immense economic value; yet, much of that value is drawn from generations of human creation: books, songs, films, images, archives, catalogues, columns and creative expertise. In other words, a portion of AI’s new wealth is built upon an older and deeper wealth created by artists, authors, performers, designers and cultural workers. Yet, in too many instances, that value is being captured without clear consent, fair compensation or meaningful transparency. Honourable colleagues, that is what is truly at stake here.

When we trivialize the extraction of cultural value in the long term, we undermine creation itself, and a country that weakens its creators impoverishes itself far beyond what any balance sheet will show. Such a country jeopardizes its ability to imagine, its ability to innovate and, eventually, its cultural sovereignty. When used as a tool, AI can clearly support human creativity, but when used as a substitute, it calls into question the place of humans in the creative process.

In reference to a bill on the use of cultural content by AI providers, my friend, French senator Catherine Morin-Desailly, said, and I quote:

By making no demands and accepting a lack of transparency, culture is engineering its own extinction by AI.

Senators, her statement is an important wake-up call because we are not talking about sudden extinction, but a gradual, almost imperceptible disappearance as the symbolic value of human creation gets lost in a flood of standardized content.

Can we really say that a work generated without human intent has the same cultural value as one born from the imagination of an artist or creator? If the answer is no, then, sooner or later, we need to ask ourselves another even more troubling question: Does our legal framework adequately protect and recognize the status and place of artists in our society and in an ecosystem dominated by AI? I have my doubts.

The culture that inspires, entertains and defines us is built on the work of creators. Their work belongs to them. It’s protected by laws and governed by a licensing market that regulates its use.

Artificial intelligence is now upsetting this balance. AI systems are trained on massive quantities of data, including protected works such as books, images, music and films. Unlike occasional human use, this constitutes industrial, systematic and often invisible exploitation that, for many creators, amounts to plundering. In the majority of cases, creators are neither consulted, nor remunerated, nor informed.

Furthermore, as advocated by the Coalition for the Diversity of Cultural Expressions, the CDCE, and the recent report by Heritage Canada entitled “Impacts of Artificial Intelligence on the Creative Industries,” Canada’s future AI strategy must uphold three fundamental principles: authorization, remuneration and transparency.

Authorization, because a creator must be able to choose whether their work can be used to train an AI. This choice lies at the very heart of intellectual property. Without choice, it’s no longer sharing or innovation; it’s appropriation. Plundering artists’ works without consent is an exploitative practice that has no place in a society governed by the rule of law.

Remuneration is the second principle, because if a work helps create value, even indirectly, its author must benefit from it. AI must not become a mechanism that captures cultural wealth without ever redistributing it.

The third and final principle is transparency, because without transparency, there can be neither trust nor accountability. Creators need to know whether their works are being used, how they are being used and under what conditions

Honourable senators, addressing these issues is all the more urgent as new problems are emerging. AI technologies now make it possible to generate deepfakes, that is, to imitate an artist’s style and reproduce a voice or image with unsettling realism. This goes beyond a mere economic issue. This is a matter of moral rights, specifically the right to the work’s integrity and the right to the artist’s identity. When AI imitates a style, reproduces a voice or misappropriates an image without consent, it severs the fundamental link between a person and what represents them.

Until recently, the meeting place between an artwork and its audience relied on institutions such as bookstores, concert halls, radio stations and broadcasters. Today, that encounter often depends on a line of code. It is no longer publishers, cultural programmers, critics or arts organizations only that guide public attention towards artworks. It is also recommendation systems designed to maximize clicks, screen time and retention.

So the question we must ask is this: Who now decides what deserves to be seen, heard or read?

As UNESCO observed in its 2018 report entitled Culture, Platforms and Machines, when the circulation of cultural works is essentially in the hands of a few dominant platforms, and when their algorithms reward uniformity, the diversity of cultural expression begins to erode. The capacity to discover unfamiliar works slowly disappears. And a culture that cannot be found is a culture at risk. In the age of artificial intelligence, the very existence of a work increasingly depends on its discoverability.

This diversity is also reflected in language. In that respect, AI offers some exciting possibilities. It can write, translate, make recommendations and summarize, but in which language?

Australian researchers estimate that over 90% of AI training is based on English-language data, specifically American English data. When a single language and culture dominate the data, they also dominate the responses, references and nuances.

The risk to the French language and its variants in Canada is clear: The language will remain present on the surface but will be absent at a deeper level. French will become a language of translation rather than a language of creation.

How can we ensure that the French language, in all of its diversity, is genuinely represented on AI platforms? That can be done by investing in high-quality corpora, supporting AI research in French, including francophone minority communities and Quebec in consultation and data training processes and leveraging the government’s purchasing power to demand truly effective French-language tools.

AI can either serve as a tool for linguistic vitality, both for the French language and Indigenous languages, or it can become a force for assimilation. That choice is ours to make as a society.

As artificial intelligence advances, not all participants enter this new era on equal footing. Large institutions have legal teams, innovation budgets and access to technical expertise. Independent artists, small publishers, regional professional theatre companies and community museums have far fewer resources.

The risk is, therefore, twofold: A technological divide can quickly become a cultural divide. If only the best-financed can create, produce and be visible through AI, then the next generation of talent may disappear before it is ever seen or heard.

The government’s role and responsibility in this regard are clear: ensure a plurality of voices in AI. This requires robust support programs for small organizations, affordable access to tools and training, shared resources in terms of data and expertise, and targeted support for independent creators and under-represented communities.

It also means that, when negotiating licences with major platforms, AI must be developed, trained and operated in a way that preserves, represents and strengthens cultural diversity without imposing a single cultural norm or unduly exploiting shared heritage. There must also be transparent mechanisms for oversight, correction and value-sharing.

Honourable senators, Canada is a recognized leader in AI. Our researchers, universities and companies are helping to advance this field, and that is something our country can be proud of.

Now that AI exists and is advancing at breakneck speed, the real question is this: What role will we play in governing it? If we do not make our own rules, we will import the rules of others. If we do not defend our languages, our markets and our creators, other priorities will take their place.

Canada has a unique strength: It knows how to balance innovation, diversity, bilingualism, cultural pluralism and the rule of law. We must therefore fully incorporate culture, official languages and Indigenous languages into any future Canadian strategy addressing artificial intelligence, not on the sidelines, but at the centre of things.

Canada does not have to choose between technological leadership and cultural leadership. In the 21st century, the two are inextricably linked.

Let’s preserve cultural diversity and also profit from the untold opportunities that AI has to offer.

During the National Summit on Artificial Intelligence and Culture held in Banff, which I had the pleasure of attending last month, three verbs took centre stage: build, empower and protect. Building trustworthy innovation, empowering creative talent and protecting what binds us together: our languages, our creations and our collective imagination.

Properly managed, AI can expand access to creation, enhance the capacity of our creators, supercharge our use of knowledge and advance knowledge.

French writer, politician and intellectual André Malraux said that “Art is the shortest distance between two people.” Let’s ensure that AI serves rather than replaces this path. Let’s follow the example of the Haudenosaunee people, who teach us to think seven generations ahead before we act.

In response to AI, this wisdom has never been more timely. The choice is ours. Let’s make sure that it’s a choice worthy of future generations.

Thank you. Meegwetch.