THE STANDING SENATE COMMITTEE ON HUMAN RIGHTS
EVIDENCE
OTTAWA, Monday, April 27, 2026
The Standing Senate Committee on Human Rights met with videoconference this day at 5 p.m. [ET] to examine and report on the impact of artificial intelligence on human rights and economic security in Canada, especially in relation to vulnerable groups and the international human right to work; and, in camera, for consideration of a draft agenda (future business).
Senator Paulette Senior (Chair) in the chair.
[English]
The Chair: Good afternoon. I’d like to begin by acknowledging that the land on which we gather is on the traditional, ancestral and unceded territory of the Algonquin Anishinaabe Nation.
My name is Paulette Senior. I’m a senator from Ontario, and I chair this committee.
I now invite senators to introduce themselves.
Senator Robinson: Good afternoon. I’m Mary Robinson, senator representing Prince Edward Island.
Senator McCallum: Mary Jane McCallum, Treaty 10 territory, Manitoba region.
Senator Karetak-Lindell: Nancy Karetak-Lindell, Nunavut.
Senator Arnold: Dawn Arnold, New Brunswick.
Senator K. Wells: Kristopher Wells, Alberta, Treaty 6 territory.
Senator Pate: Welcome. I am Kim Pate. I live here on the unceded, unsurrendered, unreturned territory of the Algonquin Anishinaabe Nation.
The Chair: Welcome, senators and all those following our deliberations. Today, our committee will be continuing its study on the impact of artificial intelligence on human rights and economic security in Canada, especially in relation to vulnerable groups and the international human right to work.
This afternoon, we will have two panels. In each panel, we will hear from the witnesses, and then senators around the stable will participate in a question-and-answer session.
I will now introduce our first witnesses, who have been asked to make opening statements of five minutes each. With us in person, from Farm Credit Canada, is Mohamad Yaghi, Vice President, Innovation Hub and AgExpert. Joining us by video conference, from Amnesty Tech, is David Nolan, Senior Investigative Researcher.
I welcome you both, and I now invite Mr. Yaghi to make his presentation, followed by Mr. Nolan.
Mohamad Yaghi, Vice President, Farm Credit Canada: Thank you, Madam Chair and honourable senators, for this opportunity.
Farm Credit Canada is Canada’s largest agricultural lender, serving over 100,000 producers, agribusinesses and agri-food operators. We are a commercial Crown corporation mandated to support the long-term viability of Canadian agriculture.
Canadian agriculture is facing a productivity problem. The OECD report Agricultural Policy Monitoring and Evaluation 2025 is direct: Canada’s agricultural total factor productivity is well below the world average. The Bank of Canada has called Canada’s broader productivity gap a break-the-glass moment. Agriculture is not exempt from that warning.
Agriculture cannot hire its way out. Extension services have thinned, agronomist visits have become less frequent and experienced operators have retired, taking decades of irreplaceable knowledge with them. Farmers are being asked to make increasingly complex decisions with less access to the expertise those decisions require while costs rise and climate volatility intensifies. That is the vacuum artificial intelligence is uniquely positioned to fill, not by replacing people but by filling a gap that was already there.
From Farm Credit Canada’s experience financing the industry, serving customers in every province and investing and building agricultural technology, I want to offer three perspectives.
First, AI earns trust when it is grounded in sector data. AgExpert, Farm Credit Canada’s farm management software, is used by over 28,000 Canadian producers managing 7.6 million acres of farmland across Canada. That data is Canadian and agricultural. It is owned by the farmers who generate it. Farm Credit Canada does not collect or use it for our own financial or lending purposes.
A joint study with Statistics Canada found AgExpert users see a 7% increase in farm productivity upon using the platform, but the number that stays with me is what that data makes possible beyond the balance sheet. One farmer recorded every field activity and input decision for years. When he retired, his granddaughter returned to the farm and learned her family’s land through that structured history. She now makes better decisions than either of them could alone. That is not automation but a story of continuity that the industry will continue to encounter. That is exactly the knowledge transfer Canadian agriculture has been losing every time an experienced operator walks off the land for the last time.
Second, the most valuable application of AI is not automating tasks. It is democratizing decisions. Farm Credit Canada’s AgExpert is developing a farm-level digital modelling capability that lets producers simulate outcomes before committing capital. Right now, that kind of analysis depends on whether you can afford a consultant. AI removes that barrier. The farmer with 200 acres in northern Ontario gets the same quality of thinking as the agribusiness with 20,000 acres in Saskatchewan. The economic benefit is not just efficiency at the farm level. The OECD shows that AI can lift labour productivity growth by up to 1% annually, and that gain comes from giving producers better decisions before capital is committed.
Third, access cannot be an afterthought. Only 50% of First Nation reserve areas have broadband meeting the national standard, compared to 93% of the rest of Canada. Connectivity is a prerequisite. For communities that have historically had limited access to agricultural advisory services and formal knowledge networks, digital tools matter even more. Where connectivity exists, AI is already acting as an equalizer by giving Indigenous communities, and all other communities, unfettered access to trusted agricultural knowledge through tools such as Root AI, Farm Credit Canada’s large language model.
On labour displacement, Statistics Canada’s 2024 research found only 6% of AI-adopting Canadian businesses reduced employment because of AI. More fundamentally, the jobs AI is filling in agriculture are not jobs that currently exist. They are jobs that were lost years ago, quietly, without a headline: the extension officer whose position was cut and the agronomist who stopped making farm calls. AI is not a threat to the right to work in agriculture. Used correctly, it is an expression of it.
In 2024, Canadian agriculture contributed $149 billion to GDP, employed 2.3 million Canadians, and set an export record of $100 billion. What it has lacked is the infrastructure to make every producer as capable as the best producer. Farm Credit Canada is committed to building that technical foundation because Canadian agriculture has what the world needs, and AI ensures every Canadian producer has what it takes to deliver it.
The model is replicable: Start with the data the sector already generates, build collective intelligence from it and distribute that intelligence equally. Those are not agricultural principles. They are the conditions under which AI raises the floor for everyone.
Thank you, and I welcome your questions.
The Chair: Thank you, Mr. Yaghi.
David Nolan, Senior Investigative Researcher, Amnesty Tech: Honourable senators, thank you for the opportunity to be here today and appear before the Standing Senate Committee on Human Rights.
My name is David Nolan. I’m a researcher on a team known as the Algorithmic Accountability Lab at Amnesty International. Our team examines the adoption of artificial intelligence technologies by states around the world and the impact this has on human rights.
We understand governments are keen to harness opportunities presented by AI technologies. However, I wish to start with a short preface that discussing any potential benefits or harms of AI is complex and multi-faceted.
First, we must acknowledge AI is a nebulous term. It is an umbrella phrase used to describe an array of varied technologies, ranging from chatbots to predictive analytics tools.
Second, any discussion of AI must go beyond the technology itself and consider the political, social, economic and cultural incentives and environmental factors that give rise to its development. This makes discussing the human rights risks and harms wide-ranging.
Therefore, I will first present select cases of how AI is impacting human rights, followed by a brief discussion of AI as an industry, and I will finish with some proposed next steps.
To date, Amnesty International has worked on a number of AI issues, with a predominant focus on the International Covenant on Civil and Political Rights, or ICCPR; and economic, cultural, and social rights, or ESCR, frameworks. We have documented how AI systems have restricted the freedom to peacefully protest through the now commonplace deployment of facial recognition technologies by law enforcement agencies around the world.
Second, AI systems have reinforced exclusion, discrimination and bias across society. A wealth of evidence has been produced in the past decade documenting how algorithmic decision‑making tools have automated bias against marginalized groups. These systems have introduced harmful repercussions across numerous areas of life for these people, including social protection, healthcare, immigration and criminal justice.
Third, AI systems have undermined the right to social protection. For example, in Sweden, an algorithmic system used to detect benefits fraud disproportionately flagged women, individuals with foreign backgrounds and low-income earners for further investigation. This system was discontinued as of last year after an investigation by the Swedish Authority for Privacy Protection.
In relation to labour rights, understandably much of the conversation is focused on the potential for job market disruption and the risks of workplace automation. However, the evidence on this remains mixed.
There are three other vectors that I wish to discuss here. First, platform labour, often known as the gig economy, is a key example of how the incursion of new technologies, such as algorithmic management tools, into existing or emerging labour markets often serves as a way of facilitating the exploitation of already vulnerable workers. Second, AI tools used in hiring and recruitment processes, including CV screening, stand to make discriminatory decisions, reproducing existing imbalances in hiring practices. Third, AI systems are often built on invisible or hidden labour, usually performed by precarious or otherwise vulnerable workers, often referred to as “ghost workers,” in the supply chain. This refers to data annotators who are key to training and maintaining the AI systems these companies are using, who often work for third-party contractors in countries or contexts with weak legislative frameworks around labour laws.
Additionally, beyond AI’s applications, we cannot ignore the extractive power and practices underlying AI’s expansion and development in this moment. AI development raises serious environmental and social concerns. These systems rely on computational power, driving the expansion of data centres that require large quantities of drinking water for cooling and threatening local communities, particularly in water-stressed regions.
At the same time, market power is increasingly concentrated in a handful of big tech firms, which is not only a competition issue but a profound human rights problem. The dominance of major technology companies undermines privacy, access to information, freedom of opinion, workers’ rights and freedom from discrimination.
We are at a critical juncture. The current trajectory of AI poses a fundamental threat to human rights and society, with the harm often falling on those most marginalized. We must challenge the false dichotomy often presented between regulation and innovation and work to build a new vision for technology, including legislation and norms, to ensure new technologies are rights respecting.
While we do not call for any particular model of AI regulation, we ask that it be grounded in a human rights-based approach by accounting for the intersectional harms of technologies. We ask that any proposed legislation prioritize binding and enforceable, rather than voluntary, commitments by companies. We ask that impacted communities be meaningfully included in policy‑making processes.
In terms of where AI and automation relate to government, we ask you to critically assess whether the automation and deployment of AI is the correct and most appropriate approach to reaching public policy or other stated aims. We also ask for clear red lines on the development and deployment of AI that is incompatible with human rights violations, for instance, banning the development, production, sale and use of biometric technologies and AI systems that enable mass surveillance.
Thank you for your time.
The Chair: Thank you, witnesses, for your statements. We’ll go to questions now.
Senator Robinson: My question is for our Farm Credit Canada representative, Mr. Yaghi.
We’ve heard testimony on AI and robotics, and I want to get clarity there with regard to agriculture. Could you give us your definition of “AI”?
Mr. Yaghi: Thank you for the question.
We would define “artificial intelligence” as the use of statistical pattern recognition at scale to help individuals find insights, make decisions and automate processes. They can use tools like machine learning and large language models to come to those conclusions, but, ultimately, it is making better use of information than before. It is really about using foundational data to increase insights or the other examples I mentioned.
When it comes to agriculture, there’s an important distinction between automation and what we’re seeing now with artificial intelligence. When it comes to robotics, it’s more about a physical labour task that’s repetitive as opposed to what we’re seeing more of, which is the conversation around artificial intelligence focused on augmenting human intelligence. That’s what Canadian agricultural needs more of, and it’s what we’re building today.
Senator Robinson: You mentioned succession. In Canada, it speaks volumes as the average age continues to rise. It means we’re not attracting younger entrants enough to sway that average from going anywhere but up. I am looking at what I see as a looming crisis on our horizon. You probably know the number of dollars that are set to transfer in the next 10 years. We know that Canadian farmers manage something like $1 trillion in assets. When we look at how we attract and retain youth in agriculture, it’s a hugely risky industry to be involved in. I think you mentioned a farmer who had tracked everything, and that enabled his granddaughter to come in and run the farm with all that knowledge. You talked about efficiency and improvements in productivity.
Can you give us a bit more of a sense, through that productivity lens, of what it actually means for a farm — I forget the number you used — and how it improves productivity and our competitiveness?
Mr. Yaghi: Thank you for the question, senator.
When it comes to productivity, the definition I would use, especially in an agricultural context, is “the value of outputs to inputs, land price and labour.” It is that trade-off between the two. It really comes in the form of total factor productivity, especially in agriculture. That’s the metric we use to define the output of agriculture in Canada, especially primary agriculture.
In regard to the land transition that we’re seeing in the sector at the moment, over the next 10 years, we’re going to be seeing one of the biggest transfers of assets from one generation to the next. At the moment, the average age of a producer is 58, so we’re going to see a lot of changes over the next few years.
We’re operating in an environment where the agricultural sector, especially primary agriculture, is facing chronic labour shortages. I believe the Canadian Agricultural Human Resource Council, which you used to lead, suggests there are over 60,000 job shortages in the workforce at the moment. Those are roles that artificial intelligence and the tools associated with it can help fill. It will never fill a human role because artificial intelligence and the tools associated with it work best when augmenting and expediting human potential as opposed to just operating alone.
In terms of the labour shortages that we see, in 2022, a shortage of over 28,000 jobs in the agricultural sector contributed to a loss of over $3.5 billion. Artificial intelligence and the tools associated with it, as I mentioned in my definition, can help augment and fill those labour shortages we see in the sector currently. It helps us in Canada to increase the competitive nature of the industry, nationally and economically, because agriculture today contributes to over 7% of our economy; it is one in nine jobs. When we compare ourselves to the Americans, for instance, they’re already assessing that artificial intelligence and the tools associated with it don’t decrease employment; they actually increase employment as well as wages. That’s a similar trend we’re seeing, broadly, across the economy: Over 83% of new jobs over the next 10 years will require post-secondary education. It is just enabling us to go into roles with higher wages and giving us more opportunities in terms of productivity output that the OECD suggests we should have.
The Chair: Can I poke at that a little bit with respect to labour and what transitions you would expect to see? For example, Canada has relied significantly on farm workers from most of the Global South. Do you see any impacts with respect to the introduction of AI and the kinds of tasks that would be completed by AI versus the need for ongoing workers from parts of the Global South?
Mr. Yaghi: Thank you so much for the question.
Again, in regard to labour productivity and the Temporary Foreign Worker Program we currently have in Canada, the view of artificial intelligence that I find is more prevalent today is focused on augmenting human intelligence. In terms of just those physical tasks and labour, that’s what we are seeing more in robotics.
That’s the division between artificial intelligence and robotics that I want to emphasize. Robotics will focus on physical tasks, as opposed to artificial intelligence, which will help producers understand insights and increase their analytical understanding of their farms.
One thing to really take note of is that, when it comes to agriculture, farmers take on some of the most risk in our economy today. If you look at soil alone, two thirds of our biodiversity is contained in our soil, so you’re dealing with an immense amount of risk constantly. There will be changes, of course, to our labour workforce over the next 10 years, but in terms of where I see the most development happening, it is going to be through the augmentation of human intelligence that these tools will bring to the table.
Senator K. Wells: My question is for Mr. Nolan. Based on what you were sharing with us, do you have any specific concerns related to Canada right now?
Mr. Nolan: Thank you so much for your question.
Regarding Canada, my understanding is that there is currently a lot of focus on the right to privacy with respect to artificial intelligence. While the right to privacy and surveillance risks introduced by intelligence systems are cross-cutting issues, as we were just discussing, with the wide-ranging or umbrella term of AI, surveillance and the right to privacy cut across all of this.
However, there are other notable concerns, and these are often contextual. My role focuses on government use of artificial intelligence systems around the world. This is with respect to what would be referred to as narrow AI systems, used for welfare distribution, within criminal justice settings and in other areas like migration, with, for instance, visa screening tools.
With respect to Canada, Amnesty International would like to see the human rights risks posed by technologies introduced in these settings taken into account. Any kind of government automation, if it’s to be introduced, should be introduced with meaningful transparency and the necessary accountability mechanisms to have the correct checks and balances when governments are trying to automate parts of public administration —
Senator K. Wells: Thank you for that. I know Amnesty International sometimes has specific reports or fact sheets related to countries and AI. If there’s anything specific to Canada, we would welcome that.
In terms of legislation at a national level, is there model legislation out there that you’re aware of that includes robust human rights protections related to AI?
Mr. Nolan: Thank you for your question. Just to circle back to your last point, admittedly, we have not done specific work relating to Canada through Amnesty Tech, our technology and human rights program, over the past few years.
In terms of legislation, we have conducted advocacy on the EU Artificial Intelligence Act over the past five years, and this provides a model example of some kind of horizontal legislation. We currently take issue with that specific legislation. However, it does provide a model in a global landscape where regulation is firmly off the table at the moment.
As I mentioned at the end of my statement, we look for certain principles of legislation, as opposed to specific models, because we realize it’s a brand new set of technologies, and there are a series of reasons why introducing regulation and legislation is difficult within the context of how AI systems are deployed across a variety of sectors and contexts. Obviously, the risks and human rights harms are often context specific.
The EU example does provide one model. Obviously, with the recent developments in the EU — I’m not sure how closely you’re following the recent omnibus proposal — some of these protections are being watered down, around transparency in particular. That’s around the registration of high-risk systems in the public domain, and this is a key ask for us as civil society and us as an international human rights organization. Transparency is a necessary first step, and registering algorithmic systems used by governments around the world is absolutely crucial to this, so it’s fundamental that this is within the regulation.
It’s also important that deployers of artificial intelligence systems aren’t allowed to decide or mark their own homework as to whether their systems should be deemed to be in a high-risk category. Thank you.
Senator K. Wells: You mentioned the principles you recommend rather than specific legislation. If you could follow up and send us a link or a copy of those principles, that would be very helpful.
Mr. Nolan: I’m more than happy to provide that after this meeting.
Senator K. Wells: Thank you.
Senator Pate: Thank you to both witnesses.
Following up on previous questions, I’m interested in safeguards, whether or not they are through regulatory schemes. I heard you mention, Mr. Nolan, suggesting some principled approaches. What are some of the risks if we don’t get this right in the human rights context?
Mr. Yaghi, what are some of the risks in terms of human rights regarding the lack of protection for not only the farming community but also workers who might be there?
Of course, please expand on anything you’d like regarding the benefits and how those are promoted in a regulatory framework.
Mr. Yaghi: Thank you so much for the question.
When it comes to the way we view artificial intelligence and the tools associated with it, it’s very much along the lens of how to create equalization. When we look across the landscape, our mandate as a Crown corporation is to bring equality across the sector so that we can increase productivity. When we look abroad at different initiatives started in different jurisdictions, like the United States or Brazil, we see, for instance, that the American Journal of Agricultural Economics found that wages and employment increased, so our mandate is to see how we can use those tools and provide them, at a base, to our producers so that they can leverage them to be globally competitive.
When we look at our stance compared to other jurisdictions, I think it’s a really foundational moment right now for us to ensure that Canadian producers and the overall economy have access to these tools in order to take advantage of the economic output. Last week, a study from Statistics Canada found that for businesses that adopt AI processes in their line of work, they actually have a productivity impact of 17%. If you factor cost in it, it’s over 5%. There’s an incredible amount of potential in these tools. Of course, with any tool, there have to be guardrails, but I believe that this committee can bring a balance between the human rights focus and the question of how we can enable this industry to benefit from the economic productivity impact of these tools. And it comes in many different forms.
Mr. Nolan: Thank you so much, senator, for the question. In terms of some of the specific principles that we discuss when we talk about introducing legislation, some of those were mentioned in the latter remarks of my opening statement. In particular, we ask that there be a consultation process and clear, easy-to-access, transparent and accountable policy-making processes that enable meaningful and equal participation for a wide range of rights holders. In particular, we want participation from impacted communities, those actually subject to the systems in the first place. In relation to algorithmic management tools introduced in workplaces, we asked that workers and trade unions are consulted. It’s absolutely crucial in centring policy discussions around the needs and priorities of those communities, and enable equal participation of representatives, advocates and organizations.
In terms of other specific principles, we do ask for clear red lines in the development and deployment of AI that is fundamentally incompatible with human rights law. We ask for the banning of the development, production, sale and use of biometric technologies by all public and private actors that enable mass surveillance and discriminatory targeted surveillance. We equally ask for a ban on the development, production, sale and use of AI systems that create and expand facial recognition databases through the untargeted scraping of facial images from the internet and CCTV footage. We also ask for the ban of automated risk assessment and profiling systems in the context of migration and a ban on the use of artificial intelligence systems in what is known as predictive policing by law enforcement. That’s the prediction of crimes by individual persons and/or in given spaces and/or in given times. One final critical point that I want to mention is that we ask that any regulation avoid loopholes and blanket exemptions regarding law enforcement, national security or military defence for the development and deployment of AI technologies. Thank you.
Senator Pate: Thank you. Almost 20 years ago, I was asked to attend, at Stanford in California, an iRobot conference, where one of the plans being laid out was to potentially use only electronic surveillance of prisoners.
I was asked to come, probably because of some of the work I had been doing and because a number of AI-generating individuals and researchers were concerned about the potential human rights impact. To follow up on Senator Wells’s question, do you have any additional examples? You gave some around social assistance and other areas. Do you have other examples of where these kinds of technologies have been used to disproportionately target certain groups?
Mr. Nolan: Thank you for the question. Yes, most of the examples I provided previously were in relation to the distribution of social protection schemes. That is where much of our work has been focused over the past three to four years, particularly in the EU. There are numerous examples of — first, within the criminal justice system — systems used to predict the rate of re-offence and therefore make criminal justice decisions. There is a notable investigation from 2016 led by ProPublica looking at a tool called COMPAS on which I am happy to provide more information, but it was discriminating against people in communities of colour.
There are other examples.
The Chair: I’m sorry to cut in. We seem to be having some difficulty with the interpreters. I think there is an issue, possibly, with the sound quality so it is causing some difficulty with interpretation. I’m pretty sure other senators would like to ask you questions. Can we utilize an approach where folks are able to ask their questions and then, Mr. Nolan, you would be able to provide responses in writing?
Mr. Nolan: Yes, I am happy to do that.
The Chair: You can still stay for the conversation.
Mr. Nolan: Yes. Absolutely.
The Chair: Thank you.
Senator Arnold: I have two unrelated questions. Mr. Nolan, I was very concerned when you said that the EU is watering down some of its transparency regulations. Through all of our hearings, everyone has said transparency is so important. They have also often said that the EU is doing things right. What, specifically, are they doing that is concerning to you? And I have a totally different question for Mr. Yaghi, concerning education. We’ve heard that Canadians are not embracing AI very robustly. Do you have any good examples of situations where people have learned about it and got on board?
Mr. Yaghi: Thank you so much for the question. One example that comes to mind is just looking across the country when it comes to rural revitalization as an issue that’s reoccurring without the agricultural sector. It’s not necessarily about studying artificial intelligence. It’s about, as an agronomist or accountant or lawyer, how can I leverage these tools to help me in my line of work?
As we’ve seen shifting demographics in the country, especially within our rural communities, we have seen tools that utilize artificial intelligence, like large language models and even predictive analytics. The beauty of it is that anybody of any background can study these tools and leverage them immediately. That’s what I discuss: the democratizing effect of it. Some of the most beautiful examples I’ve seen across the country include an accountant being able to help more producers with their bookkeeping because they can leverage these tools right now: The same accountant is able to leverage our platform AgExpert to help producers understand their financial situation but also help them predict the future of their farm. When it comes to agriculture, one thing I would love to express to the committee is that there are so many risks that producers and operators across the country have to deal with. It’s unlike a factory floor, where you can control that environment. You’re constantly trying to negotiate with nature, and in that pursuit, these tools that they are leveraging enable them to manage that risk. I speak from personal experience.
When it comes to that education aspect, a lot of folks that we see entering the agricultural sector have to leverage these tools right now, not only because they are able to assist more, but in terms of competition and productivity nationally, at an economic level. It is not about folks not using these tools and then trying to operate in that environment. It’s more that if you are not leveraging these tools, you will be left behind, so how do you adopt them into your work?
Previously, I mentioned that artificial intelligence tools are strongest when they amplify the strengths and advantages that folks have already. So if you are a producer who thinks about yields, how can you use your structured data to have a better understanding of your farming operation? If you are an accountant, how can you use artificial intelligence with your bookkeeping to help even more clients so you can automate processes as well? It is not about replacing labour; it is about exemplifying and amplifying your skills to a broader community and increasing your market share as well.
Senator Arnold: Thank you.
The Chair: Mr. Nolan, I’m not sure your audio has been fixed, so we will accept a written response, as we requested.
Mr. Nolan: Yes. Absolutely.
Senator McCallum: In the United Nations report Governing AI for Humanity, experts were very concerned about harms related to the societal implications of AI: 78% regarding damage to information integrity; 74% regarding inequalities, such as concentration of wealth and power in a few hands; and 67% regarding discrimination and disenfranchisement, particularly among marginalized communities. As for intentional use of AI that harms others, 75% were concerned regarding use in armed conflict by state actors; 72% regarding malicious use by non‑state actors; and 65% regarding use by state actors to harm individuals.
They said:
Risk management requires going beyond listing or prioritizing risks, however. Framing risks based on vulnerabilities can shift the focus of policy agendas from the “what” of each risk (e.g. “risk to safety”) to “who” is at risk and “where”, as well as who should be accountable in each case.
Can both of you comment on that?
Mr. Yaghi: I can comment to an extent. Again, when it comes to our position as a Crown corporation, it is really about focusing on how we can equalize access to different digital tools within the agricultural sector across the supply chain. When it comes to the tools that we provide currently in the market, AgExpert, for instance, or Root AI — which is a large language model that utilizes trusted and curated sources of data to provide extension service-level knowledge to producers across the country — we’re helping fill a gap. It is a gap that needs to be addressed because, from what we hear from our producers — again, digital tools can help, but when it comes to what we’ve seen, advisory services are more impactful. How can we leverage those advisory services to exemplify the knowledge that we collectively have as a country and to share more freely to producers as well? That includes producers who might not have a legacy within the sector. When I say they might not have a history, it is for a variety of different reasons.
Again, it is about levelling the playing field for everybody and sharing the best-in-class knowledge with our producers across the country.
Again, as we see input costs rise, the questions that we hear across the country are just magnifying the issue in terms of how we can continue to increase productivity and continue to feed Canada and the world. These tools are able to help us answer those questions.
Again, there will always be nuance. Artificial intelligence is not the be-all and end-all. It is not the golden key to every problem we have, but it will definitely help us when it comes to productivity and focusing on those issues.
Senator McCallum: Thank you. Is there any other information that you want to share with us that we haven’t addressed yet?
Mr. Yaghi: Absolutely, and thank you for that question. When it comes to our work within Indigenous agriculture, of course, there is an unfortunate legacy that we have as a nation. But with respect to the way we’re trying to equalize the playing field, especially with First Nations and Indigenous communities across the country, at the moment, the median farmer income for someone who is not Indigenous is around is $140,000 a year. The median income for a First Nation or Indigenous farmer is around $69,400. There is a huge variance here — of $75,000, basically.
These tools we are introducing to the market can help. Again, it is not going to be the be-all and end-all answer, but it will help close the gap. According to our own research at Farm Credit Canada, we do see helping fill that gap to be a $1.5-billion opportunity. Leveraging these tools to help all communities across Canada can definitely help our ecosystem but also help our economy.
When it comes to disenfranchised and minority workers, it comes back to how we can make that access to knowledge and information accessible. According to the OECD, the reason Canadian agriculture’s total factor productivity has been declining is discontinued investments in innovation and knowledge, and knowledge includes advisory services. That’s our attempt to help fill the void there. Going back to my opening statement, artificial intelligence is helping us plug that gap we currently see in the market.
Senator Robinson: I want to talk about the human resource. You mentioned the Canadian Agricultural Human Resource Council. In 2022, their research revealed there were 28,000 unfilled positions in agriculture. We know that cost primary producers $3.5 billion in lost sales. I want people to think about that more: how risky it is for a producer to take on the investment involved in agriculture without that kind of certainty of labour, because that means you don’t have the opportunity to earn the income to pay for your investment. And we know most of the vacancies are filled by international workers, despite robust attempts to hire Canadians. Obviously, labour is a risk in agriculture.
I have to keep reminding myself of the difference between robotics and AI. In your opening remarks, I think you said AI was not taking jobs but rather doing jobs in agriculture that stopped existing a long time ago. Could you elaborate on that a bit?
Mr. Yaghi: Thank you so much for the question. Going back to that study that saw a $3.5-billion productivity loss because we weren’t able to fill those vacancies, when it comes to jobs that were lost, we are seeing a lot of changes in the industry right now. Over the past few years, input costs have risen. When I say “input,” that means fertilizers, fuel and labour. It includes the fact that our rural communities are changing. Jobs that were once there, because of population decline, aren’t there any more.
How are we able to continue providing services across the country but fill in the void that has been left? Again, these tools are able to help. They are not going to be the ultimate solution, but they are able to fill in the gaps around knowledge and advisory services.
Again, going back to the few studies I mentioned, with the OECD focusing on total factor productivity — again, the factor used to study productivity in Canada — and seeing the sector’s performance is below average compared to the rest of the world, the reason we see that is because of a decline in innovation and knowledge services accessible to producers across the country.
When it comes to ways that we are able to exemplify the sector, it is really by providing these tools, whether it is a through a machine-learning tool, predictive analytics, building out neural networks — even through sorting the grade of crops, just through visual images, basically. They are able to detect the quality of grains, potatoes, et cetera, just through different images. By studying all these images frequently, we are able to save time. The potential of these technologies will enable Canadian agriculture to fill in the void that has been left behind by many different changes and rising challenges.
In addition, when we talk about climate volatility these days, using this information will enable operators across the country to manage their risks far more effectively. As I mentioned a few moments ago, two thirds of the world’s biodiversity is just in our soil alone. How do you manage that risk without having all the information at your fingertips?
Again, I’m not trying to create a rosy story. I’m trying to create a factual one where these data points will enable us to manage risk far better. Of course, at the end of the day, guardrails must be established, but in terms of filling the void that we saw with the lack of productivity, and because of these labour shortages, this will help us fill the gap to an extent.
I’m really confident about the next few years ahead of us in terms of collectively establishing a mandate where we can adopt these tools more broadly across the sector but help our producers increase their productivity by giving them more certainty about how they can control and manage their labour as well.
Senator Robinson: We know AI can be scary, but we are also learning that some good things are going to come from AI. But I think we’ve all gone through the shock and awe of some of the AI stories, so I am happy to hear some good-news stories.
Mr. Yaghi: I am happy to share them.
If I can share one thing with the committee — I believe AI has become a bit of a clichéd term. What I mean by that is that it encapsulates a lot of different processes and tools. In terms of the different innovations I see across Canada and the different investments that Farm Credit Canada is making in ag tech, we are seeing a huge rise of promise within the industry but also for helping our national economy as well. So thank you for the question.
The Chair: I would like to ask a question that may also apply to you, Mr. Nolan, but certainly to Mr. Yaghi. Mr. Nolan, if you have any input, I would appreciate that as well.
Mr. Yaghi, you mentioned connectivity, and I wasn’t clear on the context in which you mentioned it, because I’m thinking rural, remote and even northern communities. I’m also thinking of Indigenous communities. I would like to understand the issue of connectivity as it relates to agriculture.
You also said that AI helps to democratize decisions. I would like to know how that works.
You then talked about AI acting as an equalizer, but you also talked about this gap in terms of the earnings of Indigenous folks compared to non-Indigenous folks. I am curious as to how AI could narrow that gap. We have heard in previous testimony that AI can also really reinforce discrimination and bias. I would like to understand how AI would be able to do that within the context that you’ve shared with us.
Mr. Yaghi: I love that question. I will try to be as quick as I can and not take an hour with my response.
Starting with connectivity, in order to use tools that are online, you need access to the internet. Unfortunately, there are certain communities in our nation where broadband connectivity is around 50%. That’s unfortunate, because access to these tools really enables them to understand and get these services at the tips of their fingers. However, when we think about accessibility, we design our products with that in mind. For instance, with AgExpert, we have a desktop version, so they don’t need constant connectivity to still have access to the technology. It is very much a design-first orientation when it comes to thinking about connectivity and understanding the landscape and how we can attract new entrants into the sector. Also, for folks who don’t have that connectivity, we’re able to provide them that service regardless of whether they have an internet connection.
When it comes to democratization, it is an important point because the way we look at information we hold today, of course, data privacy is paramount. We need to ensure that no data is used against our users. At the same time, if we are collecting data points across the country, our philosophy is this: How can we enable it to help other producers? For instance, when it comes to adopting new practices — I will get a bit technical — if I wanted to adopt cover crops in southwestern Ontario, it is a viable investment for me. However, it requires capital but also a few years in order to see the upturn in yields of my farm. If you wanted to adopt cover crops in Saskatchewan, it is a far different story. You are not able to.
However, when we see these practices constructed across the country, there are information silos at the moment. The data we have right now, within AgExpert, for instance, and in other platforms as well, enables us to remove those silos — to use an agricultural reference. We are able to share that information and knowledge across the country, so practices that may be beneficial in Nova Scotia could potentially be adopted in British Columbia in certain ecosystems. We are able to share those insights across the country.
When it comes to being an equalizer, it is using that data foundation we use to build the models. Of course, there will always be concern about bias. We have built processes in to ensure that that bias is reduced by the highest degree possible, but there will sometimes be that bias, especially unintentional bias. Sometimes, there is a bit of data where, again, it might be more focused on a certain area. We have to be careful about the geographic distribution of the information we have.
However, ultimately, we are able to share that knowledge within the platform with anyone who wants access to it. For instance, over the next few weeks and months ahead, we will see more tools that will be developed, even from Farm Credit Canada’s side with AgExpert, where producers can log in to the platform and ask the tool a question about whether they should make a capital investment over the next six months. It will give a series of indicators to assess the viability of doing so.
There is a famous statistician who said, “. . . models are always wrong but some are useful.” That’s the philosophy we have in mind. It is about giving them that knowledge about what has worked before —
The Chair: Thank you.
Mr. Yaghi: I’m sorry. I tried to be quick.
The Chair: We appreciate that. Mr. Nolan, if you have any insight into this, particularly as you are focused on Europe — I can imagine that some of these issues exist there as well — we would appreciate your response.
Senator Pate: I would like to follow up on this and talk with you both — Mr. Nolan in writing and Mr. Yaghi here — about the intersections of AI with decision making that impacts democratic institutions, capitalist enterprises and how to mitigate against large monopolistic measures. I’m thinking of Monsanto. I’m looking at Senator McCallum and Senator Karetak-Lindell, because when you talk about Indigenous and northern communities, there is a great deal of work that has been done on providing climate-friendly and earth-friendly approaches to agriculture, to sustainability and to the farming of animals.
I am not sure if you have looked into those areas. What are the best approaches you’ve seen? What are the most egalitarian approaches you’ve seen? Where are the threats to democracy that we should be watching out for?
Mr. Yaghi: Thank you so much for the question. From our end, when we are looking at the application of these tools, our mandate as a Crown corporation is to be a catalyst for the industry, to ensure that producers of all sizes have access to the right technology but also the right capabilities.
When it comes to the information we’re able to share, I will go back to that equalization comment I made before about being able to make that information available broadly across the country, especially to northern communities. Let’s say there is a new entrant into the sector, but they don’t know what to invest their time in, specifically with a practice or a crop rotation. That’s not fundamentally fair to them because they don’t have those years of experience. How do they get that knowledge?
Again, there is nuance. There is never a golden solution, but we have done a lot of work behind the scenes to build these models so that when it comes to that knowledge and sharing that knowledge as widely as possible, it is accessible to anybody in the country, whether they are a 20,000-acre farm or a 200-acre farm in northern Ontario.
Giving them that viability to have a better understanding of what to invest their time and money into reduces risk significantly. I return to the examples I shared about risk factors a farm has to consider. I would argue that farmers in Canada have the highest risk appetite because every season is a risky season, especially with climate volatility these days.
How can we provide that information to them? That’s our role as a Crown corporation: to provide those tools.
The Chair: Mr. Nolan, we look forward to your response as well.
Senator McCallum: I want to go to environmental stewardship. How do you practise environmental stewardship with respect to the impacts of agriculture? I’m looking at fertilizers, herbicides and the creation of blue-green algae. On top of that, AI uses a lot of water to cool its machines down. I understand it has to be clean; it cannot be treated water. Can you comment on that?
Mr. Yaghi: Thank you so much for the question. There are a few important distinctions to make here.
I will flip the question, actually, because when it comes to the use of inputs like fertilizers, farmers don’t want to spend more money on fertilizers. They want to spend the least amount possible. We provide avenues to them through our program, which is, basically, how to apply fertilizer in the right amount and in the right way at the right time. We provide them with that information so they don’t have to use a lot of fertilizer.
When you look at Canada’s soil alone, it stores over 25 years of man-made emissions.
Carbon is a fundamental component of soil — NPK and organic C — and organic carbon is fundamental to any plant life. That plant life is essential for livestock.
We do see decreases in emissions in the sector, which is encouraging, but the use of these tools will actually enable producers and operators across the country to be able to make the best decisions. Ultimately, when it comes to fertilizer application, they don’t want to spend more; they want to spend less. When it comes to promoting organic carbon in their soil, that actually captures carbon from the air and brings it into the soil — and I’m sure Senator Black is having many discussions about this at this moment.
When it comes to the environmental component, there’s a lot going on. The way we’re able to use this technology will actually have a net-positive impact on our overall national climate inventory.
Senator McCallum: What about water usage?
Mr. Yaghi: I don’t know much about the water usage of these data centres. I never have access to them, so I can’t comment at the moment. I would be happy to look into it for you.
The Chair: I believe that was mentioned by Mr. Nolan, so a response to that would be helpful.
It seems we have come to the end of our questions for our first panel. Thank you both for your thoughtful presentations and answers.
I will now introduce our second panel. Our witnesses have each been asked to make an opening statement of five minutes. These will be followed by questions from senators.
With us at the table, from Knockri, we welcome Jahanzaib Ansari, CEO; and, via video conference, all the way from Taiwan, please welcome Audrey Tang, Taiwan Cyber Ambassador. Thank you both for being here. I will now invite Mr. Ansari to make his presentation.
Jahanzaib Ansari, Chief Executive Officer, Knockri: Thank you for having me. I greatly appreciate the opportunity, chair and honourable senators.
First, let me acknowledge and thank you for your service to the nation. It is, indeed, an honour for me to be here today. I thank you for the opportunity to appear before you to discuss the importance of AI on human rights and economic security in Canada, especially as it relates to susceptible groups and the international human right to work.
My name is Jahanzaib Ansari. I am the founder and CEO of Knockri, a Canadian organization that has been operating for a decade. Our mandate is very simple: to ensure that every Canadian has a fair shot and equal opportunity, and that no one is ever held back from opportunities they rightfully deserve by unfair gatekeepers.
Our platform is used by large and sophisticated organizations across the private and the public sectors to support hiring, promotions and training. Our solution is an AI-based one, as I was sharing, which we started about 10 years ago now. Collectively, the organizations that use our solution represent a workforce of about 3 million individuals, spanning from Canada to Tajikistan and all the way to Thailand. We’ve been covering quite a large span of a workforce for 10 years now. It’s extremely interesting how I started this company.
About 12 years ago, I was applying to jobs and just wouldn’t hear back from a lot of employers. I have a long ethnic name, “Jahanzaib,” and I wouldn’t hear back. I was very frustrated at that juncture. So I anglicized my name. I went from Jason to Jordan to Jacob, and literally, in six weeks, I got a job.
I’m not here to gain any sympathy, but I feel there are so many skilled Canadians who are probably being overlooked, and there has to be a better solution to address biases and also create efficiencies in the hiring, promotion and training processes. So, I came together with a machine-learning scientist and an industrial organizational psychologist to essentially create Knockri.
What we have created is a solution that helps organizations reduce bias in hiring, increase diversity on the basis of merit, improve efficiency and also deliver a great candidate experience. And, unlike some of the other AI hype at the moment, we started our company from the ground up with a lot of guardrails in mind and have really ensured that every single decision can be traced to how it was actually made.
So we have essentially created a glass box of an algorithm that is transparent, auditable and builds a lot of trust. My cofounder is Faisal Ahmed, a machine-learning scientist from the University of Toronto; and our chief scientist is David Mayers, who specializes in selections and assessments.
Esteemed senators, today, of course, AI has the potential to either reinforce existing inequities or help reduce them. From my observations, spanning over a decade of working with our customers across the private and the public sectors, I would just like to make four brief points, and then we can open up the floor to any questions.
I have to admit that AI is increasingly connected to the international human right to work, and that is, of course, what we are currently seeing out in the market right now, along with some of the challenges that come along with that. But when AI is trained responsibly, it can actually uphold a lot of human rights, including the international right to work.
Through a lot of our customers, when it comes to the private and public sectors and the Armed Forces, we feel that transparency is key and accountability is non-negotiable. So, building out solutions that are auditable and that you can actually open up and truly understand how a decision is being made is tremendously important.
Lastly, a lot of this is directly tied to economic security and our national competitiveness. There is, of course, a challenge of bias in algorithms; however, at the same time, we need to have regulation that will not slow us down to the extent that adoption is not there anymore and we’re lagging behind.
In conclusion, of course, AI is moving at a dizzying pace right now, but we have seen similar transitions in the past. If guided responsibly, AI can expand opportunities, reduce biases, strengthen economic security and, of course, uphold human rights.
I thank you again for the opportunity to be here, and I look forward to your questions.
The Chair: Thank you, Mr. Ansari.
Audrey Tang, Taiwan Cyber Ambassador, as an individual: Chair, deputy chair, honourable senators, thank you for the invitation to appear before you in my individual capacity. My name is Audrey Tang, and my perspective today is shaped by my current work as Taiwan’s Cyber Ambassador, my fellowship at the Oxford Institute for Ethics in AI and my experience as Taiwan’s first Minister of Digital Affairs.
During my tenure in day-to-day government, our mandate was not just to make people trust technology; it was to make digital institutions worthy of people’s trust.
Your study asks how AI affects human rights, economic security, vulnerable groups and the international right to work. I would like to offer one frame: AI is more than automation. It redistributes attention, authority and bargaining power, and in a democracy, such redistribution must be visible, contestable and co-governed.
Taiwan’s success against AI-generated scam ads showed that democracies need not choose between technocratic control and platform inaction. Citizens deliberated on the balance between fraud prevention and freedom of expression, and the same principle applies to work. Affected people should help set the rules before the systems harden into infrastructure.
The right to work in the age of AI must include three practical rights: the right to learn, the right to know and the right to contest.
The right to learn means training before displacement, not after. Work is more than income: It is apprenticeship, belonging, care and dignity.
The right to know means that when AI affects hiring, scheduling, promotion, benefits, education or public services, people should know that it is being used, who is accountable and whose data is shaping the decision. A black box decision should not be treated as due process.
The right to contest means affected people can challenge outcomes without needing a degree in computer science. Appeals must lead to repairs, like corrections, compensation, policy changes or retiring the system altogether.
This matters most for people already made vulnerable by existing systems. I’m thinking of Indigenous communities, migrant workers, people with disabilities, children, seniors, racialized communities and those underrepresented in labour and skills data. AI must not become a new way to extract knowledge without consent, to score people without context or to make exclusion more efficient.
At Oxford, my work in Civic AI translates the ethics of care into six governance questions: Are we hearing those closest to harm? Is someone named and accountable? Does the system work in context? Do those affected have recourse? Does it build solidarity rather than vendor lock-in? And does it know when to stop?
For high-impact AI, democracies should require decision traces, independent audits, accessible appeals, public incident reporting, worker and community co-governance, sunset clauses and procurement rules that avoid lock-in.
A democratic system must be interruptible, possible to pause, override or retire without disrupting essential services people depend on.
Inclusive prosperity is also democratic security. Canada and Taiwan are both free and open societies, and we know our adversaries are testing our seams of trust, but sovereignty is not solitude. It’s the way to protect people and co-operate without surrendering public judgment.
To the familiar agenda of protecting, empowering and building, I would like to add one verb: co-governing.
Protect people from harm, empower them with skills and knowledge, build trustworthy public infrastructure and co-govern AI with workers, families, communities and future generations who will live with those consequences.
A good enough ancestor does not ask whether the machines are ready to replace humans; a good enough ancestor asks how machines can help humans care for one another, deliberate together and keep faith with those not yet born.
No one should be automated out of agency. No community’s knowledge, language or labour should be treated as raw material without consent. No worker should have to negotiate alone with a black box.
Thank you. I welcome your questions.
The Chair: Thank you so much, both of you, for your very engaging presentations. I’m sure there will be a lot of questions for you.
Senator Arnold: Thank you both for being here. It has been really interesting.
Audrey, I listened to you on the podcast “Wild,” and I think that the hardest job you have here today is synthesizing everything you’ve done, because you’ve done a lot.
First, have you written a book?
Audrey Tang: Yes. The book is public domain, freely available online at Plurality.net.
Senator Arnold: Awesome, thank you.
You really live the co-governing model. I’m wondering if you could give us a concrete example in Taiwan that you implemented to really co-govern around AI.
Audrey Tang: Certainly.
As I previously mentioned, in 2024, we convened what is called an alignment assembly to respond to the harms in generative AI caused by scammers and fraudulent ads online. As people know, “deepfakes” in 2024 were very prevalent in all democracies, but as Asia’s most free — with respect to internet freedom — country, Taiwan simply cannot do top‑down censorship. Therefore, we sent SMS text messages to 200,000 random numbers around Taiwan, asking what we should do together. People chimed in with their ideas, and we chose 447 people in a mini-public, statistically the same as the wider policy, in tables of 10. They deliberated. The only simple rule is that AI only facilitates, and they have to convince the nine other people at the same table before their idea bubbles up.
Long story short, we implemented a set of ideas that more than 85% of people agreed with in the mini-public and that the other 15% can live with. Those ideas include joint liability, mandatory “know your customer” and slowing down connections for foreign platforms that do not adhere to our liability rules. Throughout 2025, impersonation or “deepfake” ads were down by more than 90%. I think this conclusively showed that when people want to show up at a table, the idea is not to do top-down control but rather to invent a bigger table.
Thank you.
Senator Arnold: Thank you. It just takes a lot longer, right?
Audrey Tang: It takes a long afternoon. It’s what is called a deliberative poll. It runs as long as a rigorous poll runs, which is usually a day or a few days.
Senator Arnold: Second, I thought you did such a good job of explaining the differences between big AI and specific forms of AI — I don’t know what the terms are, “general-use AI,” perhaps, and then the specific ones. From an energy use perspective, could you describe those to us, please? This comes up over and over.
Audrey Tang: Certainly.
Currently, in AI training, what is called a general-purpose large model needs to anticipate pretty much every use, from folding our proteins to folding our laundry, in the same model. In doing so, it’s incredibly energy inefficient to train. However, when we know what we want the model to do, for example, folding the proteins or folding the laundry, then we can train what is called a domain-specific model or a local model that incorporates the local community’s input in such a way that it also protects their data from extraction to the clouds of foreign big tech companies.
The one idea I will share with you is that with the extractive part — the very energy-consuming part — you can think of its data as oil. This kind of extraction goes to some large refinery somewhere, but the local way to train the small models — we can think of the data as soil. The local community tends to it together, fine-tunes it and continuously trains it so that whenever there’s a bias or an error, the course correction is immediate instead of waiting for the energy-consuming run that would take half a year or something.
Senator Arnold: Thank you.
Senator Robinson: Audrey may have answered some of my questions, but I wanted to pose one to Mr. Ansari.
You mentioned four points. You mentioned that when AI is trained responsibly, it can uphold standards, I think. You also mentioned transparency and accountability being key.
I was wondering if you could expand on those and tell us how those two things apply within your business. How do you ensure that AI is being trained responsibly? How do you ensure you have transparency and accountability?
Mr. Ansari: First, thank you so much for the question.
I will share at a high level, because I’m not super technical; however, I will provide you with a good understanding of the solution.
At our organization, we are not utilizing any large language models, so we have a lot of autonomy as to how to train the data. When we started the company, we had objectively taken a look at a competency. What does “growth mindset” mean, very objectively? What does “collaboration” mean? What does “agility” mean? Based on those, taking the exact definitions, we mapped the world of work with the success that correlates in the workforce.
So, if I’m going through the process, it’s not being trained on historical and biased data. A lot of the vendors out there would just see very generally — and it’s happened quite a lot in the industry — and they have trained the models off historical data. If there are existing biases in that, they will leak into your algorithm, which will create further data.
We have not taken a look at historical data. As an example, if members of a subgroup at an organization — let’s say South Asian and male — are the highest performers in these tech jobs and we have trained the algorithm just based on that, that will create a lot of challenges when you assess White females or any other group.
We learned extremely early on that we can’t actually do it in that manner. That is number one.
Number two is that we have a diverse set of raters, as well, from a lot of different backgrounds and with many areas of knowledge who have scored these candidates, as well, just to verify if there’s validity among the two.
We’ve had to conduct a lot of studies with universities and various organizations because it is a very litigious area in terms of trying to support any decisions like this. Then, when it comes to the federal public sector, it is going through the wringer of the algorithmic impact assessment and the privacy impact assessment and really taking a look at how these decisions are being derived.
Going through a lengthier process with the Innovation, Science and Economic Development, or ISED, team and then the Privy Council Office, or PCO, team and then the Public Service Commission, or PSC, team — it has been years of just ensuring that it is auditable. That’s number one. It’s very transparent. Second, it is trained in a very objective fashion, as well. I hope that’s helpful.
Senator Pate: This is for both the witnesses. Could you talk about the way you emphasize transparency and democratic participation while also remaining cognizant of sharing and protecting human rights information, particularly in a jurisdiction like ours where there are 13 or 14 jurisdictional components? I think that is different from Taiwan, but I will happily stand corrected.
Mr. Ansari: My knowledge is not extremely deep when it comes to some of the jurisdictions. However, if I were to give you an example, we had a situation where we had a woman from the Black community. She was being overlooked for the longest time. She was in line for a promotion. She was just stuck there for about five years. They brought us in as part of a process and, after five years, she actually ended up getting promoted. She moved up, and it showed us that when you build a solution that levels the playing field and assesses individuals based on merit, it upholds a lot of basic human rights, like that of having fair and equal opportunity and advancing in a career in a dignified fashion.
On top of that, senator, we don’t hold any information. We would delete all the data because it is not helpful for us anyway. Our function is to ask how we can help these individuals have a fair shot and an equal opportunity and, at the same time, give them the agency to not be tied down by an organization storing any of their data.
Audrey Tang: If I may, I would like to make a distinction between data coalitions, that is, people pooling data in a way that is useful to all of the members, versus the aggregation of data. It is possible for multiple players, stakeholders or communities to join a data coalition without sharing any of the raw data. There exists a kind of technology called zero-knowledge technology that allows people to prove that they can do something, or that they are in possession of certain knowledge, or that this community can respond to a certain kind of query, all without revealing any of the personal, identifiable information underneath.
During the pandemic, in Taiwan, we used a privacy-preserving contact-tracing method. Basically, a venue prints a random number on the QR code on the front door, a person scans it and sends it to a well-known number, 1922, but the telecom knows nothing about what this random number means, and the venue learns nothing — not even the phone number of the visitor — and the state learns nothing whatsoever. However, if an infection happens, we can perform contact tracing and use recursive notification, again, without sacrificing any of the privacy of the people who are not in the affected area.
I hope this illustration shows a little bit of the flavour of how a zero-knowledge data-knowledge sharing arrangement can actually work.
Senator McCallum: Thank you to the presenters for your work. In the 2024 UN report Governing AI for Humanity, it states:
There is, today, a global governance deficit with respect to AI. Despite much discussion of ethics and principles, the patchwork of norms and institutions is still nascent and full of gaps. Accountability is often notable for its absence, including for deploying non-explainable AI systems that impact others. . . .
It further states:
The development, deployment and use of such a technology cannot be left to the whims of markets alone. National governments and regional organizations will be crucial, but the very nature of the technology itself – transboundary in structure and application – necessitates a global approach. . . .
They outlined guiding principles, saying:
These principles acknowledge that AI governance does not take place in a vacuum, that international law, especially international human rights law, applies in relation to AI.
Can you both speak to that?
Mr. Ansari: I can start from my experience, of course, and then I will let my esteemed colleague take it from there.
Senator McCallum, what we’ve seen in the interim of a standardized approach to regulation internationally when it comes to human rights is that a lot of the large organizations actually have ethics boards. They have these committees where every single solution is deeply analyzed and assessed. If there is any chance that it will create an adverse impact to any subgroup, that solution does not advance.
From my purview, I’ve yet to see something globally that has been adopted, of course, but a lot of the organizations have very specific and concrete mandates. As I was saying, it is very litigious. A lot of companies have been sued already because these algorithms have created biases. That’s number one.
Number two is that, as it comes to Canada, as part of the federal mandate, we, as vendors or other algorithms, have to go through these algorithmic impact assessments. It is based on risk criteria regarding how critical your solution is when it comes to making a decision.
If it is used as part of the air force and you are targeting a building, for example, that’s extremely high-risk. That decision process has to be extremely well understood because it can create a lot of challenges. Our solution is very low-risk because we are just doing decision support; there will always be a human reviewing it at the end of the day.
Regarding what I’ve witnessed internationally, there isn’t set guidance at the moment. The EU has an act. Some of the American states have acts; however, Trump was trying to get rid of those, so we’ll see how that plays out.
I would say that, right now, the best plan of action, at least in Canada, is working with folks like Mark Schaan from ISED, understanding some of the regulations they are setting out and making sure that every vendor is actually legitimate.
Audrey Tang: If I may, I think the report’s diagnosis is right that a patchwork of principles without enforceable duties would not govern AI, which is intrinsically a global phenomenon in technology. But I think democracies, including and especially middle powers, should build interoperable governance, so it is not identical governance that applies everywhere the same globally, but rather auditing standards; instant reporting standards; provenance for synthetic media, as we just mentioned; and procurement requirements that avoid any kind of vendor lock-in. All these are like stacks that can be made to work across jurisdictions without harmonizing every domestic rule. I think Taiwan and Canada, as free and open societies, can be peers in that work.
The global governance deficit, I believe, will not be closed by another universal principle, but can be closed by enforceable duties that make principles contestable in each and every domain.
Senator McCallum: Thank you.
The Chair: I’m sitting here completely blown away by both of your responses to questions. I don’t think we’ve heard such responses in the past, so thank you so much.
My question for you, Mr. Ansari, is this: In light of the solutions to bias-free decisions that Knockri is working on, I’m wondering what the take-up is on the work or the benefits that you offer? We do not need their names, but what would a typical client look like in terms of the work you are doing?
Mr. Ansari: As it regards our solution, I would say federal government departments, the air force and a lot of the Armed Forces. Then, in the private sector, it is large organizations, such as tech companies, consulting companies, banks, those in the insurance sector, those in the education sector, et cetera.
The value, of course, varies from organization to organization. Some organizations have a mandate of increasing gender and racial diversity in short lists of candidates, of course. It is a very specific focus over there on how we can take a look at, first, the pipeline of talent and truly understand why certain individuals are not being screened through. Because what we have seen is that, similar to my experience, while human intelligence is great, there are challenges there as well. We are not perfect creatures.
So I would say our solution is used to increase gender and racial diversity in shortlisted candidates. A lot of organizations are utilizing it for efficiencies now. When you apply for a job, it takes so long for some people to hear back. This allows a lot of recruiters and HR teams to focus on the higher-value work of in‑person conversations, ideally, and to let the tedious work, where bias can actually creep in, be handled by automation at that stage.
That’s what we have seen with a lot of our customers. They’ve been able to save a significant amount of time and cost as it comes to that.
Recently, chair, I would say the solution is being utilized by the air force to create training efficiencies.
As I was saying, it is directly tied to economic security and national competitiveness. We are seeing that Canada is facing a lot of challenges when it comes to productivity and efficiency, and we all know that.
In addition, if I may make this point, a lot of young Canadians are leaving the country, which is not a good sign. As an example, one of my childhood friends left for Florida. There are highly qualified individuals who are leaving the country, but we also have very qualified individuals who are here locally as well who don’t have the right jobs.
So we’re seeing it now being utilized to upskill individuals. It is being utilized in universities to upskill students and to ensure they are matched with the correct competencies that the organizations are looking for. Their path to employment can, therefore, be a lot faster as well.
There are a wide variety of things that it touches. Of course, as it comes to adoption, ChatGPT has accelerated it tremendously. With adoption, initially, there was a lot of fear around the technology, and then education came on some of the challenges and biases. Now I feel as if a lot of the adoption is actually happening.
The Chair: Thank you so much.
Audrey, I don’t have a lot of time left, but I’m really intrigued by the governance model you talked about. You mentioned the importance of the people affected helping to set the rules, for example. You also mentioned that it is possible to use AI in a way that is not about replacing but helping us to relate with each other and help each other. Could you talk a little bit about that as well? I’m also interested in the take-up of the approach that you are using, whether that be in Taiwan or other jurisdictions.
Audrey Tang: Thank you. I would like to first make a distinction regarding an AI that automates intelligence — sometimes called “authoritarian intelligence” — that makes decisions on behalf of people. For example, 10 years ago on social media, many people felt their agency had been taken away, because previously, when we followed the same people, we would see the same feed. But 10 years ago, it was replaced by a very judgmental AI that personalized our feed and encouraged “engagement through enragement.” That’s very authoritarian.
In Taiwan, we call it “assistive intelligence,” a different kind of AI that assists the cross-conversation between people who would otherwise not agree. Instead of a wildfire that consumes the oxygen between people, think of it as a campfire that people sit around: It illuminates the faces of people who are different from you. Still, each campfire is tended by a bounded set of people, like 10 people, or 100 people and so on for a larger bonfire.
In Taiwan, we have designed this kind of prosocial media, such as Polis, which is an open-source technology that is being used by, I think, more than a dozen countries worldwide, including Canada. The idea is, instead of making the outrage viral, Polis makes the overlap viral. In order to be viral on this prosocial media — it highlights only the ideas that people who would otherwise never agree, agree on. So only the bridge‑makers gain virality, and in doing so, people heal their polarization and so on.
Our demonstration was successful enough that even traditional social media, like x.com — previously Twitter — have now adopted a similar algorithm called Community Notes, which lets people who bridge across different ideologies write in there to add context to any viral misinformation or disinformation or just contentious information.
Now we work with all major social media companies on Community Notes implementation and also on collaborative notes, which are notes drafted by AI and then instantly corrected by humans so that AI can learn what can translate across communities — like between the climate justice community on one side and the biblical creation care community on the other — so they can translate across their vocabularies. I hope that illustrates a little bit of the bridging potential of such language model technologies.
The Chair: Thank you. That’s very interesting.
Senator Robinson: Thank you for joining us. Audrey, can you tell us if the Minister of Digital Affairs in Taiwan has collaborated with the Minister of Agriculture to drive digital transformation in agriculture? On the previous panel, we heard from Farm Credit Canada, and they talked about the Canadian context of AI integration in agriculture. Can you provide insights from your time as Minister of Digital Affairs on any intersection between the two ministries in Taiwan?
Audrey Tang: Definitely. We indeed worked with the new Ministry of Agriculture during my tenure. Both ministries were set up, more or less, in one year’s time. In Taiwan, of course, we also face the issue of fewer young people in the agricultural sector, and the professional know-how that they have in one particular domain does not readily translate to the other domain.
Mostly, we worked with the Ministry of Agriculture to make sure that environmental sensing was not limited to one production facility, ensuring the long-term trend of what crops to plant, how to hydrate, how to ensure the customer relationship as well as online relationship management and so on. We have a program called TCloud, where each of the small- and medium‑sized enterprises, some of which are in the agricultural sector, can choose between thousands of vendors.
The thing here is that we implement the same transparency, data portability and freedom to move between different vendors so that the data stays with the agricultural operators. That way, if one prediction model or one vendor does not fit their particular circumstances anymore, they are then free to shift to another vendor.
As an incentive for a vendor to join the program, we at one time offered vendors up to an 80% subsidy from the government to help them to bootstrap, especially the small- and medium-sized start-up vendors that produce software for the agricultural sector, to introduce them to the agricultural sector. So the state subsidy here is not on one national or regional champion, but rather to the idea of interoperability and data sovereignty and ownership in the agricultural sector, so that they can collaboratively train their sector-specific models.
Now we are seeing the idea of data coalitions being taken up by other models as well. For example, the financial sector just announced that banks and insurance companies are now using the same idea to train their sector-specific models without sacrificing their data sovereignty and ownership.
Senator Robinson: That’s fantastic.
You mentioned that in Taiwan, you have an aging producer and farmer situation. Here in Canada, we are faced with that, and we are seeing a consolidation of farms being driven by the need to have economies of scale to stay in the business, have profitability and be able to survive in it.
What are the barriers in Taiwan to young people joining agriculture? How do you see AI possibly better facilitating or attracting them? We heard from our Farm Credit Canada folks about the reduction of risk through better predictability. Can you expand on your opinion on that?
Audrey Tang: Definitely. I would say that better predictability is a large part of it. In any kind of work, including agriculture, we must consider the “ABCD” that I just mentioned: apprenticeship, belonging, the idea that we can care for our communities, as well as dignity. These are equally important.
As we just mentioned, AI systems that help intergenerational solidarity, the ability for people’s know-how in one context to transmit to another without sacrificing their local wisdom, without aggregating them as oil refineries or as soil — I think this is also very important.
Also, capital and land access, the difficulty of inheriting tacit knowledge from one generation to the next, the traditional dependence on a small number of intermediaries for market access and so on are solved not by even more consolidation but rather by small operators reaching co-operative scale through shared digital infrastructure. For example, our drone agricultural service platform brings together 90 small operators who share equipment, certify pilots and compliance records. None of them could individually afford the equipment or meet the regulatory burden, but together they reach this horizontal scaling that, previously, only large agribusinesses could. I would offer that the state’s role is not to pick national champions but to subsidize the freedom to choose. Thank you.
Senator McCallum: You can answer this question in writing if you want: Audrey, how can the federal government best support Indigenous data sovereignty?
Audrey Tang: Yes. Taiwan has 16 Indigenous nations and more than 42 language variations. So we see cultural and also transcultural — the ability to translate across culture — sovereignty as very important. When we say “sovereign AI” in Taiwan, we don’t mean just a national Taiwan model that speaks Mandarin, Taigi, Hakka and other Indigenous languages. We mean a reproducible process for the language communities to own their socio-cultural composition of the data that’s curated within those language communities, as well as alignment assemblies, ways for people to draw boundaries around how AI should enter their community, almost like a code of conduct for AI agents.
Together, these two allow each community to feel that they own their own socio-cultural determination when it comes to language model training. They also incorporate that transcultural translation capability so that when one language or culture gains a certain capability, one can readily cherry-pick it into the community if the community so wishes, but the agency and sovereignty are held by the Indigenous community, not by the top-down national commission. I hope I’m making sense. I am happy to add more in writing, but that is the general idea.
Senator McCallum: In writing would be excellent. Thank you so much.
The Chair: Do you want to add anything, Mr. Ansari?
Mr. Ansari: There is one last point I would like to add because I feel the conversation about transparency was somewhat overlooked. I want to add more with respect to that.
We’ve observed that individuals, especially from groups that are susceptible, really appreciate when you share why you’re using an AI solution because there is a lot of fear around it generally. We’ve seen that, as soon as we start to communicate that, we are helping to ensure that every single person has a fair shot and an equal opportunity, and that has changed the entire dynamic of the conversation.
So if you were to employ some AI solutions, I feel it shouldn’t just be software. You need this behavioural and human understanding to also get them to understand why it’s being done and what the benefits are. Otherwise, they’re going to be very scared of it and very untrusting of the public sector.
The Chair: Thank you very much. Thank you for that last point.
I feel hopeful after hearing from both of you. Your responses were incredible. To be hearing about how sovereignty is not solitude and about inclusive prosperity, and how to use AI in a way that actually reduces barriers and biases. I can speak for myself: I’d been hoping to hear that, so thank you very much.
This brings us to the end of our panel. Thank you both for contributing to the work we’re doing. It’s been very helpful.
(The committee continued in camera.)