Agriculture increasingly integrates the use of data-driven systems. Despite the novelty of these technologies, most digital tools are being developed and controlled by legacy actors – oligopoly firms that supply farmers with products such as chemicals. As a result, these powerful firms are accumulating massive amounts of agricultural data.
This market advantage demands an examination of existing Canadian data policy to advance an equitable distribution of its benefits and risks for agri-food actors.
Yet, there is no government legislation specific to data related to food and agriculture.
There is not even a legal definition of agricultural data, which could include information about farmland (e.g., soil data), crops and animals (e.g., fertilizer application), farm equipment (e.g., fuel consumption), and business reports (e.g., farm revenue).
We implore the federal government to design new policies on ethical data governance in agriculture and strengthen proposed data legislation, Bill C-27, because our research suggests that current and proposed laws do not adequately account for the harm that could arise from the collection and use of agricultural data.
The primary approach to data policy in Canada is to regulate privacy through the protection of personal information.
So, where does agricultural data fit within current Canadian law?
If personal information is data that renders an individual identifiable – the definition used in federal government policy- then even some anonymous or de-identified agricultural data could be considered personal.
For example, with publicly available soil type data, Google maps and information about crop types, or some other combination, it may be possible to connect “anonymous” information back to an individual farm or farmer. This presents challenges for balancing farmer expectations of de-identification or anonymity and the usefulness of data.
There are two primary data policies at the federal level, both of which outline the expectations for the collection, use, disclosure and retention of personal information: the Privacy Act for the federal government and the Personal Information Protection and Electronic Documents Act (PIPEDA) for private-sector organizations and federally regulated businesses.
Other relevant federal acts include the Library and Archives of Canada Act and the Access to Information Act, which governs how the government controls, stores, shares and publishes information.
The provinces and territories often legislate privacy and access to information (or freedom of information) together, such as British Columbia’s Freedom of Information and Protection of Privacy Act and Ontario’s act of the same name. There are additional privacy laws at the provincial level that focus on the private sector and health care.
Although none of these policies explicitly address agricultural data, they could be applied to personal information collected in the context of agriculture.
For example, PIPEDA presents a legal requirement for businesses to communicate the purpose of data collection and obtain consent before it is collected. PIPEDA would apply to a farmer if they collect personal information about current or potential customers, say to circulate a newsletter.
The Commons standing committee on industry and technology is currently studying Bill C-27, the Digital Charter Implementation Act, which proposes three new statutes: the Consumer Privacy Protection Act (CPPA), which would repeal and replace PIPEDA; the Personal Information and Data Protection Tribunal Act to assist the Office of the Privacy Commissioner of Canada in decision-making and enforcement; and the new Artificial Intelligence and Data Act.
The government has also tabled Bill C-26, An Act respecting cyber security, amending the Telecommunications Act and making consequential amendments to other Acts. That bill has been approved by a Commons committee and is awaiting final approval in the House.
However, the proposed statutes continue to focus on personal information and none specifically addresses agricultural data. The regulatory landscape thus remains ambiguous and incomplete in its ability to address many of the collection, sharing and use practices pertaining to data from farm environments.
There are potentially negative consequences of government policy focusing almost exclusively on personal information.
Much of the agricultural data collection and use takes place through farm management platforms. These are usually proprietary packages of software and hardware tools designed to support farm management decisions through the collection of data which is then used in conjunction with artificial intelligence (AI) to deliver advice about farm operations.
In the case of platforms developed by for-profit corporations, the corporation itself governs the use and control of the agricultural data via internal policies (e.g., privacy policies) and legally binding documents between private individuals and the corporation (e.g., terms-of-use agreements).
As required by law, corporate policies and legal agreements need to outline the corporation’s collection and use of personal information and communicate these terms to users.
In the corporate policies for farm management platforms, data use is addressed more than any other action (collecting data, accessing data, etc.), with a greater emphasis on the use by the platform’s company than use by the farmer user or third parties.
There are limitations to the use of corporate policies and terms of use in data governance. For one, despite growing societal concerns over privacy and lack of trust, it is by now well-documented that many users accept terms without reading legal agreements or even opening the hyperlinked policy.
This phenomenon has been attributed to a lack of understanding or capacity to read legal language and fatigue in confronting lengthy legal documents.
In addition to government and corporate data policy, voluntary codes of conduct have come to play a role in governing agricultural data.
For example, acting as a third-party standards or certification body, the non-profit Ag Data Transparent (ADT) organization currently oversees corporate uses of agricultural data, assessing them for alignment with its core principles, such as transparency of data uses.
However, ADT evaluates only those self-nominated agricultural technology providers via 11 questions such as: “What categories of data does the data platform collect?” and “Do the tech company’s agreements with the user address ownership of user data?”
While empirical studies suggest that farmers trust it to some degree, research also shows that farmers view it as a peripheral service limited in its capacity to govern data use fairly.
ADT prioritizes transparency and clarity in communicating terms in data policies, rather than regulating what data management decisions are in line with user rights.
For example, John Deere Corp. received its “seal of approval” in 2018 because its answers to the ADT assessment aligned with corporate policy. Yet, this policy gives farmers using John Deere platforms no rights of ownership over the data they input or provide.
When one looks at the existing governance tools, it is obvious that a lack of government policy for agricultural data in Canada increases the power of corporations to set the terms for its use (and potential misuse).
The negative consequences of corporate-dominated food systems on citizens and the environment are well-documented.
Ethical data governance requires new government policy. A first and essential step will be defining agricultural data, which will entail delineating the diverse types therein – including data on soil, fertilizer use, yield, and data pertaining to people. This precision can then be used to inform inquiry into which current and proposed policies may apply.
Almost certainly, ethical data governance will require expanded legal rights of access over data for farmers and farm workers. But it should not stop there. Appropriate agricultural data governance may entail policies that exceed those on data – for example, competition law to address long-standing power imbalances in agriculture.
Moreover, the federal government collects and holds its own agricultural data which could be used as a testbed for implementing innovative data governance mechanisms such as data trust and co-operatives. Based on our public scholarship with farmers, it appears there is an appetite for such arrangements.
Canada faces an opportunity to shape the future of agricultural data governance. Let’s not miss it.