On December 5, 2024, the Canadian Securities Administrators (CSA) released Staff Notice and Consultation 11-348: Applicability of Canadian Securities Laws and the Use of Artificial Intelligence Systems in Capital Markets1 (the Notice). The guidance in the Notice addresses key considerations for registrants, reporting issuers, marketplaces, and other market participants that leverage AI systems in their business practices. While the Notice does not create any new substantive legal requirements, it lays out overarching themes governing the regulation of AI and securities and provides specific guidance for market participants based on existing securities regulation.
In tandem with the Notice, the CSA also launched a public consultation to understand the potential for AI systems to impact capital markets and the potential need to adopt appropriate safeguards to maintain the integrity of the financial system. Submissions in this regard are due by March 31, 2025.
Key Takeaways
- The Notice aims to protect investors by ensuring that capital market participants in Canada remain fair, transparent and accountable for risks arising from their deployment of AI systems. AI systems can be beneficial in increasing competition and efficiency in capital markets, but only when they are deployed responsibly.
- In the deployment of AI systems, market participants should
ensure the following:
- Deploy technology in a manner compliant with existing securities regulation, as it is the activity that is conducted that is regulated, not the technology;
- Develop and implement risk management policies and practices to address the inherent risks unique to the AI systems being deployed, including understanding the AI supply chain risks, importance of data accuracy, and having a "human-in-the-loop";
- Ensure users understand the decision-making process behind AI outputs to enhance transparency and accountability;
- Accurately disclose the use of AI systems to allow investors to make informed decisions; and
- Consider that AI systems must not result in conflicted decisions that favour market participants over clients and investors.
- The CSA has provided specific guidance to various market participants regarding obligations related to AI usage with reference to existing regulation.
Recent Trends in the Regulation of AI Systems in Capital Markets
AI systems in capital markets are being used for things such as in the improvement of efficiency and accuracy of operational processes, and from a regulatory standpoint, in trade surveillance and detection of market manipulation.2
At a broad level, AI systems present ethical concerns such as a lack of clarity where legal responsibilities lie where an AI takes actions or makes decisions that would normally be sanctionable under securities laws, the potential for compromised outputs where input data is inherently biased by its users, and data privacy concerns including data persistence and the "right to be forgotten". However, a growing concern in capital markets is "AI washing", which is when organizations exaggerate or falsely advertise the extent of their usage or development of AI. The misstatement of AI usage can attract investors, but misrepresentation is expressly prohibited under securities laws in Canada. In the United States, the Securities and Exchange Commission (SEC) has previously released two enforcement actions against investment advisor firms for AI washing.3 Further, the Nova Scotia Securities Commission has also noted that they are monitoring corporations in their jurisdiction for misleading statements related to AI.4
As the regulatory framework regarding AI system deployment continues to change in Canadian capital markets, market participants should remain vigilant about their obligations related to risk mitigation, explainability, fairness and transparency, in line with existing Canadian securities laws.
Themes Related to the Use of AI Systems in Capital Markets
The CSA has laid out the following "themes" relating to the use of AI system in capital markets.
Technology & Regulation
The first theme focuses broadly on the application of securities law to technology. The CSA has emphasized that securities regulation is meant to be technology-neutral and will be applied to each issuer similarly regardless of technological proficiency. However, the legislation governing securities is largely principles-based, and issuers may be required to take different measures depending on the technology in question. As AI systems and software vary in complexity and in use, the principles underpinning certain securities laws may be interpreted in different ways depending on the market activity. For example, the deployment of an AI system used to gather information requires a different type of due diligence, both in scope and depth, as compared to an AI system that automates trade execution.
AI Governance & Oversight
The Notice highlights the importance of implementing governance and risk management practices when deploying AI systems in capital markets. Policies and procedures must account for the unique features of the AI implemented and the potential risks. As examples, the CSA proposes certain policies, including having a human intervenor that can consistently monitor the inputs and outputs of the AI system, ensuring that there is adequate training and AI literacy for those who use the outputs of the AI system, mitigating the broad technological risks associated with the use of AI such as bias and cybersecurity, and conducting detailed and routine reviews of the full ecosystem and supply chain of the AI system.
Explainability
The Notice further outlined "explainability" as an important consideration when incorporating AI into any securities-related operation. Regardless of whether a market participant utilizes AI in their business activities, they are responsible for all decisions they make and for understanding their internal controls and processes. In AI systems, output is dependent on the data upon which it is trained, and market participants need to understand and explain the rationale behind the AI's decision-making process. The CSA noted that "black-box" AI systems pose challenges to the basic tenets of certain securities laws, including transparency, accountability, record keeping, and auditability. As such, market participants need to balance their choice of AI system with explainability, so that they may continue to meet obligations under securities law.
Disclosure
The Notice also outlines the importance of disclosure. The use of AI systems poses certain material risks that investors need to be aware of; for example, investors must be able to make informed decisions about whether to invest in an issuer that utilizes AI systems in their offerings. Conversely, the CSA warns against market participants making misleading claims related to their use of AI systems. As discussed briefly above, "AI washing" is an overstatement of the use of AI and its benefits in an organization, which is expressly prohibited pursuant to securities legislation in Canada.5 This conduct can be deemed misrepresentation under securities legislation as it could lead to misinformed investment decisions.
Conflicts of Interest
The last theme outlined in the Notice centres around conflicts of interest. AI systems are susceptible to bias – in order to combat this bias, market participants must consider the explainability of the system's outputs, the bias of the data sets being used to train or finetune AI systems, potentially flawed code, and the limited expertise regarding managing conflicts of interest and AI systems. The potential for AI systems to produce conflicted results and favour particular interests over others, such as the issuer over the investor, runs afoul of current securities regulation.
CSA's Guidance for Market Participants
The CSA has also provided outlines of how existing securities regulation requirements would apply to specific market participants' use of AI. The CSA provided analysis for the following types of market participants:
- Registrants:
- Advisers and Dealers; and
- Investment Fund Managers;
- Non-investment Fund Reporting Issuers;
- Marketplaces and Marketplace Participants;
- Clearing Agencies and Matching Service Utilities;
- Trade Repositories and Derivatives Data Reporting;
- Designated Rating Organizations; and
- Designated Benchmark Administrators.
The specific guidance provided by the CSA refers to pre-existing securities regulation and reflects on how the integration of AI impacts market participants' obligations. The full list of the selected rules and guidance highlighted by the CSA can be found in the Appendix of the Notice. As an example, the CSA notes that Non-Investment Fund Reporting Issuer's usage of AI can trigger their continuous disclosure obligations per National Instrument 51-102 Continuous Disclosure Obligations.6 The CSA notes that while the integration of AI can trigger disclosure obligations, issuers must use their discretion to determine whether their use of AI is of such materiality that it necessitates disclosure. The CSA also notes that disclosure is to be specific to the issuer, and should be neutral, thorough and proportionate to the materiality of the AI system's impact on the issuer's business operations.
As another example, the CSA noted that Advisers and Dealers seeking to use AI systems cannot outsource "registerable activities" to AI. In particular, while AI may assist with meeting the issuer's "know-your-client" (KYC) obligations, the Adviser or Dealer must still engage in a "meaningful interaction" with the client. The CSA references the previously issued CSA Staff Notice 31-342 Guidance for Portfolio Managers Regarding Online Advice7 and its Companion Policy 31-103 Registration Requirements, Exemptions and Ongoing Registrant Obligations8 in providing this guidance. The CSA refers to its previous regulation to demonstrate that while the integration of AI can enhance the issuer's KYC obligations, existing regulation must still be followed.
The above examples are meant to provide context of the CSA's specific guidance for market participants and are not exhaustive. Please refer to the Notice for the entirety of the CSA's guidance for market participants.
Consultation and Requests for Comment
As noted above, the Notice also included ten consultation questions which the CSA hopes will encourage further engagement from market participants. The CSA's consultation process will help guide the evolving regulatory framework for the deployment of AI systems in Canadian capital markets and will ensure the implementation of responsible and innovative technology to market participants. While the Notice contained no new regulatory action, the CSA has indicated that any changes or amendments to the rules will be made through established policy-making processes.
Footnotes
1 Canadian Securities Administrators, "Staff Notice and Consultation 11-348: Applicability of Canadian Securities Laws and the Use of Artificial Intelligence Systems in Capital Markets" (5 December 2024), online (pdf): osc.ca/sites/default/files/2024-12/csa_20241205_11-348_artificial-intelligence-systems-capital-markets.pdf.
2 See Ontario Securities Commission, "AI in capital markets – exploring use cases in Ontario" (last accessed 17 December 2024), online: osc.ca/en/industry/artificial-intelligence/ai-capital-markets-exploring-use-cases-ontario#:~:text=The%20most%20mature%20use%20of,Supporting%20advisory%20and%20customer%20service..
3 U.S. Securities and Exchange Commission, "SEC Charges Two Investment Advisers with Making False and Misleading Statements About Their Use of Artificial Intelligence" (18 March 2024), online: sec.gov/newsroom/press-releases/2024-36.
4 Nova Scotia Securities Commission, "What is AI Washing?" (8 May 2024), online: nssc.novascotia.ca/before-you-invest/what-ai-washing.
5 For example, in Ontario, see RSO 1990, c S 5 at Part XXIII and s 126.2(1).
6 Canadian Securities Administrators, "National Instrument 51-102: Continuous Disclosure Obligations" (last updated 9 June 2023), online (pdf): osc.ca/sites/default/files/2024-04/ni_20230609_51-102_unofficial-consolidation.pdf.
7 Canadian Securities Administrators, "Staff Notice 31-342: Guidance for Portfolio Managers Regarding Online Advice" (24 September 2015), online (pdf): osc.ca/sites/default/files/pdfs/irps/csa_20150924_31-342_portfolio-managers-online-advice.pdf.
8 Canadian Securities Administrators, "Companion Policy 31-103 CP: Registration Requirements, Exemptions and Ongoing Registrant Obligations" (last updated 6 June 2022), online (pdf): osc.ca/sites/default/files/2022-11/cp_20220606_31-103_unofficial-consolidation.pdf.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.