The unreasonable effectiveness of Montreal’s AI scene
From the top-level government initiatives to the activities and energy on the ground, the more I have dived into artificial intelligence (AI) in Canada, the more I am convinced they are paving the way for the future.
When I started my role as AI specialist for the Office of the Queensland Chief Entrepreneur last year, I was tasked with developing an AI strategy in Queensland. The stats on AI in Australia were not great at the time (and still aren’t). We have seen a grand total of AU$29.9 million invested in AI by the federal government (CSIRO recently announced a further AU$19 million), we have less than 100 AI startups and our research efficiency sits at 76th in the world, indicating an absence of translation to commercial outputs. At the time of research, data from job site Indeed also revealed an AI talent gap of eight times more AI jobs than AI skilled workers in Queensland.
Charged with developing our AI strategy in Queensland, I looked for inspiration from leading examples to learn from and quickly landed on Canada. In addition to the similarities to Australia in culture, political systems, the size of the economy and, importantly, the strong backbone of academic expertise, Canada simply appeared to be doing everything right.
As the first country to release an AI strategy, Canada was arguably the first to see the promise of AI, or at least the first to realise that they had an asset of deep learning expertise that needed to be fiercely protected. Their CA$125 million Pan-Canadian AI Strategy, which aimed to link clusters of AI expertise across the country, laid the foundations for an ensuing flood of investment into AI by federal and provincial government, local industry and global giants.
The Vector Institute in Toronto alone received CA$180 million investment from the Canadian government, Ontario government, and private companies including the likes of Google, Nvidia, and Accenture. Critically, embedded in their mission statement was their aim to “work with institutions, industry, start-ups, incubators and accelerators to advance AI research and drive its application, adoption and commercialisation across Canada.” This core focus on collaboration became Canada’s silent strength.
In what seemed like a blink of an eye, an entire ecosystem of AI shot up around these core AI research institutes. As shown on the map below, there are now 650 AI startups (and growing) supported by a structure of research partners, venture capital firms, incubators and accelerators, and public organisations.
The map itself is made by Canada’s Element AI, who in the space of two years have grown from seven to 500 staff, acquired over CA$100 million in Series A funding and opened their first international offices. It is no surprise they had long been on everyone’s radars as one of the most impressive AI startups to watch.
Therefore, when Marc-Etienne, head of public policy and government relations at Element AI, magically appeared in Brisbane and invited me to visit Montreal, it took no convincing. Come December, Montreal would host not only the Neural Information Processing Systems (NeurIPS) conference, one of the most acclaimed AI conferences, but also the first G7 Multistakeholder Conference on AI, and I had every intention of being there. Some frantic work organisation and borrowing of the clothes I would need to survive a 40 degree temperature drop later, I was on my way for a week of AI in Montreal.
If I had had any concerns that the AI activities in Montreal were just an illusion, my first fifteen minutes quickly convinced me otherwise. I picked up my baggage and started following a purple line out to the entrance, where some friendly Element AI staff were there to welcome all the new arrivals for NeurIPS. I stepped out to get my Uber and the airport signage itself was lit up with AI:
The message was clear: Montreal is AI.
Attending NeurIPS on the Sunday and starting to hear about the AI advancements from around the world, as well as progress on the development of ethical frameworks for the adoption of AI, I was rapidly struck with the sense that I had stepped a good five years into the future. Presentations included avatars that were starting to seem impressively human, surreal AI-generated video game graphics from Nvidia and, later in the week, a mind-blowing talk on the use of bioelectrical signals to regenerate limbs. Meanwhile, panel discussions centred on the very real challenges of harnessing data and AI in such a way to ensure human autonomy into the future.
The feeling of being on the cutting edge only increased as the week continued. Tuesday demonstrated Montreal pioneering the way in the ethical adoption of AI, with the launch of the Montreal Declaration for Responsible AI after a solid year of consultations with the public (I’ve signed it and recommend reading it). The 2025 scenarios that citizens had been asked to consider included concepts such as diagnosis through personal digital health twins and the deliberate exploit of self-driving car algorithms to cause crashes. These led to the creation of ten principles, of which I particularly applaud Responsibility (humans must remain responsible for decisions made with AI) and Sustainable Development (AI use must ensure sustainability).
On Thursday, Element AI hosted the first G7 Multistakeholder Conference on AI, bringing together seven of the leading economies to discuss challenges and strategies for AI in society. Ed Santow, the Human Rights Commissioner for Australia, set the stage by making a solid case for approaching AI from a human rights perspective, with measures to ensure that AI not be used to undermine the rights of any individual, in particular the minorities who have been underrepresented in AI development to date. Discussions gravitated to the need to create an agreed framework for the ethical adoption of AI. Unsurprisingly, Canada led the way in this with Prime Minister Justin Trudeau announcing an open call for partners into the ‘International Panel on AI’, for which France was already on board.
Beyond these talks, I witnessed the genuine commitments to AI development that Quebec and Canada at large continue to make. Prime Minister Trudeau announced a $230 million investment, industry-matched, into the SCALE.AI supercluster, bringing their total supercluster funding to almost CA$1 billion, and later there was a pledge to teach one million children to code in the next year. Minister Bains, the Canadian Minister of Innovation, Science and Economic Development, was on site to talk to visiting data experts about how to become a trusted ‘Switzerland for data’. Then, he took to the stage to moderate the lunch panel with Foteini Agrafioti and the grandfathers of deep learning, Geoffrey Hinton and Yoshua Bengio. I can honestly say I have never been so impressed with a politician.
Finally, it seemed like Canada was doing everything in their power to provide support for startups. There was an announcement of CA$6.3 million repayable funding for six AI startups to scale, with Element AI receiving CA$5 million towards compute power. I learnt of grants that provide a 40% rebate on any full-time research staff in your company - double this if you’re Canadian!- and a 24% rebate for B2B software development in Quebec. The Canadian government is currently also updating procurement processes to enable SMEs to more readily acquire their first projects and essentially every AI startup I spoke to was collaborating with multiple research institutes.
For startups looking to scale, Canada is home to a number of excellent AI-focused accelerators like Techstars AI and NextAI. Mitacs also recently announced a new global commercialisation program with CA$7 million from government to support Canadian startups taking up residence in international hubs around the world. Last but not least, for any of those tempted to set up in Montreal, those with AI skill could be in the door with a working permit and working in as little as two weeks. Montreal International will also work with you to provide free visa assistance and access to grants to support your startup setup.
So was Montreal all that I thought it would be? Yes, and even more. Beyond the effectiveness of the strong government support, welcoming policies, open research institutes, dedicated industry interest, and startups all coming together to create a whirlwind of progress, there is something very special about Montreal; there is an irresistible friendliness in the people and a contagious buzz in the AI scene that leaves you feeling like there is nowhere else you, or your AI company, should be.
I say this with full conviction because, having gone to visit without any intention to do so, I was so thoroughly impressed that I will be moving there next year. À bientôt, Montréal!
