We read every submission from Canada’s AI task force: here’s what they said

Earlier this week, Canada’s innovation ministry shared the results from the 30-day national consultation it held late last year to inform the country’s upcoming AI strategy. The consultation included feedback from 28 members of Canada’s AI task force, appointed by AI Minister Evan Solomon, which included representatives from industries and fields from tech to academia. 

 “We must aggressively tackle the negative public sentiment on Al. If we do not overcome this, we will face a backlash and many of the investments will be lost, as well as the path to a prosperous future.”

Joelle Pineau, Cohere

Across the 32 submitted reports (some members appeared to send in more feedback than initially requested), the diagnosis was remarkably consistent: While Canada remains a leader in AI research, it is also a laggard in commercialization and currently lacks the domestic compute capacity (and capital) to adequately harness the full potential of AI.

While most agreed that Canada needs to move quickly to address these issues and ensure the country’s digital sovereignty, ideas on how to do so varied widely, as respondents shared a sometimes conflicting laundry list of suggestions.

A few common threads emerged across submissions: the Government of Canada ought to pick its spots, name AI champions, take the lead in buying more made-in-Canada AI, double down on some of its existing programs, reform others, and build up the country’s compute capacity.

All of the  AI strategy task force member submissions can be read in full here.  The official government report on those submissions was put together with a mix of AI tools; BetaKit, instead, assigned a human reporter to review all 348 pages. 

You can find a compilation of the most compelling ideas from each submission, listed in alphabetical order according to the submitter’s surname, below.

Doyin Adeyemi, University of Toronto

• Launch a comprehensive audit to determine how AI systems are being deployed across government, identify high-risk use-cases without safeguards, and gauge their impact on equality.
• Implement a framework of “no go areas” for high-risk AI, modelled on the EU AI Act.
• Enact a chatbot safety law similar to California’s SB-243 to protect youth.

Ajay Agrawal, Creative Destruction Lab

• Propose five moonshot projects that are defined by a singular objective that could help a broad selection of Canadians, could not be possible without AI, drive the creation of infrastructure that might produce wider-ranging social and economic benefits, and could not be delivered by the private sector.
• Drastically reduce healthcare wait times, the Grade 3 non-proficiency rate, the harmful impact of homelessness, or the time to detect air, maritime, or space incursions, as some potential options.
• Such initiatives would require significant collaboration and aggressive budgets based on performance, and could fuel tech advances and useful learnings that may be applicable elsewhere.

Benjamin Bergen, formerly CCI, now CVCA

• Build and maintain “as-Canadian-as-possible” compute and cloud capabilities with full legal control and domestic tech integrated across the AI infrastructure stack.
• Canada’s first AI strategy focused “heavily” on AI research and “too little” on commercialization and scaling firms—its next needs to look “where the puck is going” and focus on the AI application layer, where Canada already has some advantages.
• Push back on foreign laws that allow foreign governments to demand information about Canadian citizens.

Olivier Blais, Moov.AI

• Rebuild Canada’s digital foundation with an AI Readiness Fund to modernize public and private data infrastructure, because “AI cannot thrive on fragmented, outdated systems.”
• Focus on AI skills development for the Canadian workforce beyond just engineers.
• Canada “should not rely entirely on foreign infrastructures,” but also “cannot afford to reinvent the wheels” when it comes to building a sovereign AI value chain.

Michael Bowling, University of Alberta

• Scale the AI Chairs program because it is an “AI talent factory.”
• Fast-track visas and update study permit policies to allow internships; offer “a reliable and transparent path” to permanent residence for students graduating with AI PhDs.
• Fund more not-for-profit, focused research organizations doing frontier AI research.

Shelly Bruce, Centre for International Governance Innovation

• Use Al “to achieve a step change in Canada’s foundational cybersecurity” by buying Canadian products and incentivizing medium and large organizations to follow suit.
• Make security a key part of the country’s AI focus by setting clear expectations, adding it to CAISI’s remit, and using domestic tech to defend key AI assets.
• Collaborate with other countries by offering to host the next significant AI safety and security gathering and creating a new international secretariat.

Cari Covent, Microsoft

• Launch “AI Mission Canada” programs in strategic sectors in areas of maximal benefit, such as AI triage and diagnostic tools that help hospitals cut wait times, or AI-powered precision farming solutions.
• Reward top-performing provinces and companies by tying federal support to AI adoption and demonstrable results.
• Provide funding and an AI solution marketplace to help SMBs adopt AI and connect with trusted vendors.

Daniel Debow, Build Canada

• Defer capital gains taxes for founders, investors, and employees when they re-invest in private Canadian firms and increase capital gains tax exemptions for entrepreneurs and early staff to $25 million per business to boost risk capital and retain key talent.
• Keep AI rules “clear and lightweight”—new, complex regulations and the associated uncertainty they create will stifle Al adoption, so do not revive AIDA.
• Incentivize government to adopt AI by increasing transparency and rewarding results.

Garth Gibson, Vdura

• Renew the Pan-Canadian AI Strategy while also establishing a set of national sovereign AI compute facilities, such as three 10MW data centres to start, operated by international cloud providers and aspiring Canadian cloud companies.
• Launch a domestic workforce development program to ensure the country has enough AI software engineers and data engineers.
• Establish an AI fast-follow facility tasked with reproducing, understanding, and documenting global innovations, sharing knowledge with domestic companies, academics, and governments, and spinning out tools, datasets, talent, and companies.

Arvind Gupta, University of Toronto

• Review existing research funding programs like tri-council, CFI, and Genome Canada to see whether AI excellence criteria should be included.
• Create a new multidisciplinary Canada AI Chairs program.
• Fast-track study permits for talented international students, lure them with new AI scholarships, and create effective pathways to Canadian employment post-graduation.

Diane Gutiw, CGI Canada

• Declare boosting Canadian AI literacy “a national imperative” and fund a national strategy to achieve this.
• Simplify data governance through national standards and policies.
• Establish secure innovation sandboxes and clear guidelines for “pragmatic sovereignty.”

Adam Keating, CoLab 

• Reform Scientific Research and Experimental Development (SR&ED) and the National Research Council of Canada Industrial Research Assistance Program (NRC IRAP), and tie future research and development funding to commercial success.
• Reform BDC and Export Development Canada so that they make only indirect investments in top private funds and no longer directly back startups.
• Incentivize all major foreign AI companies to establish a presence in Canada while also supporting the growth of the country’s own domestic players.

Alex LaPlante, Royal Bank of Canada

• Develop an inclusive national AI education framework for Kindergarten to Grade 12 students, integrate AI across undergraduate programs, and scale applied, industry-linked master’s and professional programs.
• Establish “AI Centres of Excellence” in priority domains such as health, energy, financial services, justice, and education.
• Create a national framework for AI micro-credentials and an outcomes-based “AI Transition Fund” that offers targeted bursaries and tax credits to workers displaced or transitioning due to AI’s impact on the labour market.

Gail Murphy, Digital Research Alliance of Canada 

• Embrace and become a global leader in responsible AI, including by leveraging Canada’s existing strengths in math, the social sciences, and humanities.
• Establish internationally recognized, focused, pre-competitive industrial research and development centres in areas where Canada has expertise, such as photonics.
• Renew and expand existing support for fundamental AI research while also launching one or two new Canadian AI research hubs, including in BC.

David Naylor, University of Toronto

• Avoid launching new agencies and structures focused on AI—instead, create a time-limited, high-level task group to review national AI adoption and deployment strategies from elsewhere and develop a “superior integrated plan” for Canada.
• Rethink how we finance lifetime learning and reskilling, given not just the impact of AI “but the reality that ensuring generations may well have multiple careers before retirement.”
• Keep supporting Amii, Vector, and Mila, but also focus any further federal postsecondary education investments on deployment and applications.

James Neufeld, Samdesk 

• Demand commercial excellence from Canadian businesses. Grants and innovation programs are not where innovation happens—rather, they often tie up talent in “zombie” companies without commercial traction.
• Name some national AI defence champions.
• Ensure former Canadian public service leaders see a greater opportunity to contribute to domestic institutions that could benefit from their understanding. Right now, they often end up advancing the interests of foreign firms after leaving government.

Marc Etienne Ouimette, Digital Moment

• Create a Strategic Compute Cluster that may serve as a hedge against external shocks and a platform for domestic AI adoption.
• Launch a Canadian Compute and Infrastructure Initiative to support the growth of the country’s compute ecosystem.
• Pursue a broad AI adoption and diffusion strategy and secure sovereign capacity to safeguard the country’s interests. “Canada should neither resign itself to total dependence or chase unattainable supremacy.”

Taylor Owen, Centre for Media, Technology, and Democracy 

• Amend the Online Harms Act to include AI platforms and create a new Digital Safety Commission empowered to receive complaints, investigate harms, and educate Canadians.
• Amend the Consumer Privacy Protection Act to ensure Canadians know when they are interacting with AI systems and AI-generated content, require mandatory human review of consequential automated decisions, and bolster public reporting.
• Adapt existing support mechanisms and develop new programs to support the production of reliable, evidence-based information by journalists and content creators as AI slop floods the internet.

Patrick Pichette, Inovia Capital (a Cohere investor)

• Announce Canada’s intent to become the world’s lowest-cost safe data centre infrastructure for countries focused on AI and data sovereignty.
• Declare and support national AI champions, such as Toronto-based large language model developer Cohere.
• Reform Canadian pension funds to ensure that they invest at least five percent of their assets in Canadian tech priorities and next-generation industries.

Joelle Pineau, Cohere

• Foster support of open-source tech to help “break the monopoly of AI.”
• Aggressively tackle the negative public sentiment on Al. If Canada does not overcome this and ensure public trust, “we will face a backlash and many of the investments will be lost, as well as the path to a prosperous future.”
• Develop the right data governance model, such as “Data Trusts,” which allow others to use the private data of Canadian individuals and organizations (such as medical records, consumer habits, and creative works) in a safe and controlled fashion.

Ian Rae, Aptum 

• “Become a critical node” in the AI hardware supply chain by leveraging Canada’s natural resource wealth and increasing the country’s capacity to refine and process rare earth oxides for magnets and semiconductors.
• Leverage and expand Canada’s strength in energy to secure “equal-value long-term power-purchase agreements with foreign cloud providers.
• Introduce legislation that requires any international cloud companies providing mission-critical services to delegate operational authority to a Canadian entity.

Sam Ramadori, LawZero 

• Forge a “middle powers international sovereign Al coalition” that pools capital, compute, and research as an alternative to the US-China duopoly.
• Establish a Canadian In-Q-Tel for sovereign dual-use Al by creating a dedicated strategic investment fund within BOREALIS to bridge the gap between Canadian Al startups and national security procurement needs.
• Empower BOREALIS to make a small number of “moonshot bets” on local business in areas where Canada has a competitive advantage, such as AI and quantum.

Sarah Ryan, CUPE

• Work with the provinces and territories to develop a model of lifelong learning about tech and incentivize employers to provide digital skills training to workers.
• Focus training on not just AI, technical, and digital skills, but also soft skills like communication, collaboration, and problem solving, which can be “difficult if not impossible to automate.”
• Retool and expand existing federal programs and explore other models for delivering workplace training and credentials.

Sonia Sennik, Creative Destruction Lab

• Position Canada as a strategic customer and focus on “procurement over grants.”
• “Supercharge” existing networks like CDL that can identify and accelerate high-potential ventures early.
• Target high-impact industries with the largest potential return on investment from AI adoption, such as healthcare and critical minerals.

Michael Serbinis, League

• Launch a new, $2-billion pre-seed and seed-focused fund-of-funds, a $2-billion Canadian AI growth fund for Series A and B startups, and a $5-billion sovereign wealth fund targeted at growth equity companies.
• Harmonize Canada’s capital gains tax with the US and adopt a qualified small-business stock equivalent.
• Adopt the US SBIR model and mandate federal agencies to set aside a percentage of their research and development budgets for small businesses.

Louis Têtu, Coveo

• Build a shared AI stack that is “on tap,” or easily accessible to businesses and governments across Canada, and has “a sovereign switch” so no foreign companies or other countries can turn it off.
• Back homegrown AI champions as they scale domestically and abroad.
• Improve government service delivery and cut costs using AI.

Natiea Vinson, First Nations Technology Council 

• Commit $5.2 billion to close the First Nations connectivity gap by 2030.
• Dedicate a portion of Canada’s $2-billion Sovereign AI Compute Strategy to Indigenous governments and organizations so they can establish sovereign AI infrastructure, data governance frameworks, and compute access.
• Establish a federal-provincial funding stream to support AI literacy, digital skills, and workforce training for Indigenous peoples.

Mary Wells, University of Waterloo

• Take a “tiered approach” to categorizing AI risk with clear “red lines” where the risk is too severe and not well understood enough for AI to be deployed. The EU AI Act—which focuses on applications and outcomes and gauges risk from minimal to limited, high, and unacceptable—offers an example of how this might be achieved.
• Require a human in the loop for mission-critical agentic AI.
• Establish an independent AI regulator, ombudsman, or commissioner to monitor government use of AI.

Feature Image courtesy Image courtesy LinkedIn.